 Well, good evening, everyone, and thanks very much for attending the session. I know it's the last session, it's the second day, so we will try to make it as beneficial to you guys as possible. My name is Kashif Thikhar, I run Worldwide Sales, BizDev, and Services here at Plum Grid, and this panel is actually really unique, and when we were thinking about this session, we were thinking about bringing people from different parts of the world to share their experiences. That I think is one thing that is unique about this session. The other is that OpenStack is disrupting a lot of technology models, and the other thing it's disrupting is a lot of business models. How different system integrators, resellers, and also technology companies are offering services to their customers. So we'll go across the globe here and get some perspectives from this panel, and then we'll take some questions. So why don't we start off with going from left to, I guess, left to right. Please tell us a little bit about yourself, and you each have a lot of history as people, and a little bit about the company, so everybody gets a perspective, which region, what you do, and what have you been doing in the past? Hi, I'm Mike Mesco. Work for Onyx Incorporated. You're a system integrator, reseller, working the U.S. and the federal space and some commercial. My background is network admin, Linux admin, was a virtualization engineer for a while, and then when OpenStack came around, pretty much identified that as the way to go. Started in on Diablo and compiled a lot of stuff from Trunk back then, and I've kind of worked up to current, and I'm really glad to be here. Good seeing you guys in the audience. My name is Rodrigo. I was the first employee of a company called Mercado Libre that is like eBay for Latin America. So basically, at Mercado Libre we implemented OpenStack in 2011, a long time ago, we have a huge case of what we did at Mercado Libre, and I worked there for 14 years, then I decided to create a company called Novelio, where basically we help companies from South America, Latin America, to implement OpenStack and use OpenStack. So basically the company was created by myself and the whole CloudVider team from Mercado Libre. So we know OpenStack since 2010, I think, so we have a huge experience in that. Okay. My name is Ruth Harmsen. I'm founder and CEO of Fairbanks, which is a very nice name, but it's in the Netherlands, in Europe, in EMEA. I would like to thank you, Cash, to be here and to tell you a bit more about our organization. Fairbanks is in the European market in the Netherlands for about 10 years now, and in 2010, 2011, because of the problems into the market, we started with VMware to do implementations, storage implementations, and in 2010, 2011, we started on the marketing side with the Blue Ocean strategy within our management team, and there we decided to look more to open source. At the same time, we met people from Rackspace and the NASA and also from Nebula, and there we had the first contact with OpenStack, and we were, well, very, very happy to see that technology, and we decided to do more on that, and in the meantime, we did a lot of implementations already. Hi, my name is Yuji from Nesik. Arigato gozaimasu. Arigato gozaimasu. I taught Japanese to him, so my name is Yuji, anyway, from Nesik, Nesik is a system integrator and then doing the distribution business for Japanese IT market, and I was a network system engineer before coming to the US, it was two years ago, but right now, my title is a business development, so I'm looking for a good product or services for Japanese IT market, and then two years ago, I found PramGrid in the Sanibel, and then we started the four-cloud business two years ago with PramGrid, and then, still, we are a distributor, and that's all. Okay. Hello. I'm also very honored to be here. My name is Minehiko Nohara from Makunika Networks, Japan. We are quite similar, I'm Makunika Networks, the distributor. Our business model is to find the very new technology from abroad to bring it back to Japan and adapt it to Japanese market. I've been working for the virtualized technology for the past few years, and now I'm a cloud product dedicated engineer. This is a little bit of color here, so I think you see our friends from Japan here, we've been spending a lot of time in Tokyo. As you know, the next OpenStack Summit is going to be in Tokyo, and I think the markets across the world, especially in Japan and other Asian countries, also in South America and in Europe, they're seeing a huge transition. They've been using certain technologies, but I think there's a move in a desire to acquire new assets, new technologies that are very disrupting along the cloud area. Thank you very much for the introduction. Maybe Nohara-san, I start with you. If you can give us your sense, because you've been in this market and you work with a lot of customers, give us a little bit of sense from your vantage point, the state of OpenStack in the customers you meet and in Japan in general. The virtualization market is very growing in Japan at this moment. I think the data is a bit old, but at the beginning of last year, it was set about 50 percent of the companies already start virtualizing their servers. In fact, most of them are built by vSphere environment, but most of them who are using vSphere is also considering to move on to OpenStack environment because not only cost, but they want to get rid of the vendor locking. That is very serious especially for the cloud-provided people and enterprise people. I think it will change how to say. The market is changing now in the time. The next OpenStack summit is in Japan, and there are a lot of Japanese in this conference as well. I heard it from the Canonical friends. Canonical had the private party on the last Sunday, and then 20 percent of the attendee was Japanese, meaning a lot of Japanese are interested in the OpenStack. And then, actually, we, the Nesic, deployed first prom-grade deployment to our customer. It is not POC, it is not trial, it is the production use case. The company name is the NTT Smart Connect. They are doing the service provider business in the west of Japan, and then I'm honored to deploy that prom-grade and OpenStack to them. Then, here is the representative from the NTT Smart Connect. I want to introduce to you guys, he is Inoue-san. So, Inoue-san, can you show your face to the audience? And then, I think he is one of the hottest social, solution architect in Japan. I mean, he is using, he likes to use the cutting edge or the new technology for their business and for their infrastructure. And then, he is looking for the good or new technology for their environment to compete with someone. So, if you have a good idea or a good product, you can introduce your product to him, he will be happy, I think. And maybe, Inoue-san, expand on that, why do you see OpenStack in Japan? Why, because there are other options as well. Yeah, the other option for the OpenStack is the same as similar to the opinion with Nohara-san. But I think that there are two reasons. The one is cost effective and then the flexibility and the scalable. But I'm not sure we can deploy the OpenStack the cheaper than the VMware or not right now. So, we are still thinking about that thing. So, we are looking for that good solution still. That's where you come in and help. Yeah. Okay. All right, Drup. Well, if you look to OpenStack in the Netherlands, we also hear it here, the discussion about the pets and the kettles. If you look to the companies in the Netherlands, we have more pets than kettles. So, what we see is that the hosting companies and the multi-tenant companies, the media companies, they are very interested in OpenStack. And I think if you look to, I speak with a lot of CIOs. And now after the problems we had in the market and the business is going well also in Europe and also in the Netherlands, we see that the ability to bring new technology into the companies is also in the Netherlands number one. And of course, Fender Lock-In is one of the items and also the price is an issue. But I really believe that innovate new technology is one of the issues we see in our country. So, Latin America is something similar to what is happening. So, we go to customers, we go to enterprise and we typically, we see VMware, V-Block and ISPs trying to use V-Blocks and try to implement it for years. And they are locking. I mean, it's not easy to move ahead from that. And also, the functionality is very close. So, we go and say, okay, we know OpenStack. We implemented OpenStack in different companies. We have a good case that people know us from Mercadolibre and we have the team that is working for you. And typically, they like us. They sometimes are a little afraid of OpenStack because they think that OpenStack is terrible, difficult that maybe is correct. But we are building tools to help those companies to implement OpenStack. So, the first thing we did is we know that OpenStack is anybody can download OpenStack and install it, right? But it's not the same for an ISP, for a company that wants to put things at production. So, what we did is we created a tool that deploy OpenStack ready for production very fast, like 20 minutes. So, then what we see with people that the first impression of OpenStack is critical. I mean, because they used to see VMware and VMware next, next, everything works. OpenStack, you need to understand perfectly what you are doing. So, for us, the first impression is critical and we are deploying tools on top of OpenStack to help the users to use OpenStack in a way very easy, like many wizards set up your cloud. So, that is how we go to a customer and help them to see OpenStack. So, first impression is critical for us. And we are going to ISPs to enterprise and they're really like what we are doing and we are doing very good business with them. Okay. So, I want to make it a little bit interactive. If you have any questions, I can pause and take because I think you have a variety of people here from different regions, different experiences. So, any questions before I proceed? OpenStack generally is not a shy crowd. So, yeah, please. Okay. So, let's repeat the question. The question is that OpenStack requires a specific skill set, which is investment, especially from these integrators and resellers, investing in resources before they can provide services to their customers. So, in the people and the return, compared to the existing technologies that are available. Go ahead. Okay. Good question. Very good question. So, we know how much is VMware. We know also that with VMware, we don't pay only a license. We pay all the support services and many, many, many other things. So, for us, we know perfectly OpenStack from long, long time ago. So, for us, if you have the right people, people that really know OpenStack from long time, investing in people is what we are doing. It's the way to go. It's easy, especially in Latin America that is completely different from US. And so, yes, we did the cost analysis and we decided that we're gonna make tools to help users use OpenStack. And we put a lot of effort in UI, UX, many things to know how to help users to have production environments very easily. So, even that is really, really cheaper compared to VMware. In our design, we built some of the support and help you get started engineering services into kind of our capacity pricing. So, it's upfront. You realize what the comparative expense is gonna be. I've got a plan for a year to get this in production supported and then the customer is in a position to decide, do they wanna start hiring in their OpenStack expertise? How much do they need? What does their support level, day-to-day operations level look like? Or do they wanna continue using us for that as they are still figuring it out? One thing that we find customers are really looking at more these days is they had kind of an easy mode with the later versions, the more recent versions of vSphere. It's quite figured out. It works very well. With OpenStack, there's an engineering effort. There's a lot of upfront and that investment pays off at scale. You've paid a lot of money for vSphere to have a really well-done, figured-out product, but I don't need to pay for that engineering on rack number 25. Two years. But I think just to comment on that thing, that's a good point. There's always a comparison here. I think the data is also moving really fast. I think figuratively, just from a Plumbered perspective, we started here two and a half, three years ago, knocking, meant like 300, 400 customers. The data is changing rapidly every single six months. I can tell you in the last six months, the sophistication of the tools that are being built and the ease that different providers are bringing in from an OpenStack perspective, from a network virtualization perspective, it's changed quite a bit. It's shrinking significantly. But I think the clear, I mean, for example, we used to go to, Tokyo, everybody used to talk about CloudStack, but I think even if you look at that, that's a click, click, click deploy kind of a model, but that is changing, right? Because it's very clear that the community is converging certainly around OpenStack. And that's at least our humble perspective with the sliver of the customers that we meet. Anybody else want to chime in? Okay, maybe another question. I think questions are much fun. Anybody else? Well, maybe I can give an answer to your question because what we did, we started as an IT company 10 years ago, do implementations of storage, implementation of VMware, help customers to virtualize their data center. But we saw that in 2006, we were one of the earliest companies in the Netherlands who did the implementation with customers. For instance, Ctricks at that time was very difficult to virtualize. We said to each other, we can, and it was one big customer in the Netherlands, and well, I'm a salesman. So I said, we can manage that, and at the end we did. But at that time, you saw, and that was our problem within our company, because the competition was very big. So the customer can choose a lot of companies, and they can, so we saw prices for the consultancy fails or drops down, prices on hardware if you are a reseller. Earn your money also on hardware. So we saw everything increases. So we said to each other, we have to change and go to another model, go to another market. And that was our blue ocean strategy to change. And now, for a consultant, and your question was, how long does it take? For us, it was a road for two years. But now we have consultants and do the job for our customers for a much better price. And that for us brings us more. And when you are one of the first companies, and we are in the Netherlands, the community ambassador for the foundation. So it brings us a lot. That's what I was thinking, and what he was talking about. We as a company have timed the time it takes to install a fully functional open stack environment. Just maybe gives you some idea. It used to be days and probably weeks, year before last and now roughly we timed it. Literally we had a timer for about 1500 core environment. We were able to bring it up in less than an hour. So everything was automated. Everything was Ansible Publisher strips, done. So I think it's becoming much, much easier. You have to plan for it, but it's not like VMware is a piece of cake. So let's maybe talk a little bit about the specific use cases that you guys are seeing for your customers and maybe it applies to different people over here. So Mike. So we target the federal sector primarily. And what we're seeing is the adage that open source goes with public sector. It's finally really catching on. We've seen a lot of adoption of open source, but now it's really moving down to the infrastructure level. And the public sector, hosting centers, large data centers are starting to act a lot more like the large commercial hosting centers. They're in competition with them now. You can go get Amazon for Gov. That's a big competition in a market that government to government hosting didn't used to have. So there's now a huge drive and a reason to do all this innovation because you want to keep reporting to the same work. You don't wanna be sent to another job within the government. It's time for them to start betting. And we're seeing that they're starting to. They're looking at where's the innovation happening? Where's the community going? Where is the next five or 10 years of hosting gonna be? And that's OpenStack. That's pretty well figured out by the people that are obviously at an OpenStack summit. But everybody else seems to be coming around. We're definitely leaving the early adopter phase and starting to move into the early majority. And we see that in the federal space as well. And what would it be the typical, let's say, use case that is still DevOps in federal? I mean, whatever you can share. Yeah, we see DevOps type stuff. DevOps is, we won't say it's catching on in the federal space yet, but it's on the radar. The idea that there's ops guys and developer guys, it's breaking down. They're starting to get tired of throwing stuff over the fence. When the developers are definitely qualified to go and crank out a few networks with SDN via Horizon or a script and crank out a bunch of VMs and then start automating that. Ops guys should be moving and concentrating on engineering. And so as that starts to work out, that'll become a bigger use case. We've had some customers that were asking they're very forward leaning and they're talking about DevOps. Our primary customer in use case is more around just I as hosting. Like I worked for a place that's kind of just a I as hosting center government to government. And the whole deal is pay X dollars a month, get some VMs, right? They didn't really want anything beyond that. And OpenStack, a perfect fit for that. Yeah, real. So what we are seeing in Latin America, so it's e-commerce companies, they're trying to have agility. They're trying to not have the IT department as a bottleneck. So they are looking into OpenStack. So for us, e-commerce is very easy because at Nouvelle we came from an e-commerce, the huge e-commerce company in Latin America. And also we are seeing ISPs that they're trying to build a private cloud because you go to an ISP in Latin America, always they talk we have cloud and so on. But when you try to understand how the cloud is set up, it's virtualization, like old style guy that create the machine and so on and put the IP. So there's a different way to do it. We can do it for less money than an hour vendor. And we can do it very fast and we can do a POC in 20 minutes and show you what you can have with OpenStack. And also banks, we are seeing banks that they are trying to investigate OpenStack and we are trying to help them because what we see with banks is they are very structured, very close set, and we said, okay, we can help you guys to create a private cloud very easily and very fast and secure. So but the common denominator is that basically they want API driven. So they want maybe developers create the machines without using horizon or any tool, but others they want to use the horizon or what we did. So they want agility, speed and secure. So what we are seeing in Latin America? The banks, is it for quite a while or do you see it's coming up now? No, we are a new company, we're a seven month company and at the beginning when we go to the bank, they said, no, no, open source, we don't like open source, but then we show them and they're starting to see that OpenStack is only a layer and then they are still working on the data center with their own firewalls and they're starting to accept it. We have a few banks as a customers. We are talking with many banks in Latin America. Well, in the Netherlands, banks are very anxious on doing OpenStack for the moment. We really believe that the innovation on OpenStack will bring also the opportunity for banks after 2015. I really believe that. Now we see in the Netherlands, the media companies, they started early to use OpenStack because of the tremendous amount of data they were having. So combination of OpenStack and for instance, SEV brought that the media companies were looking for OpenStack at that time. Yeah, I wanted to talk about him. Yeah, I wanted to talk about the media, but he said, so I need to change the topics, but of course, service provider is the... Yeah, that's fine. Completely same, so the reason, but almost the service provider in Japan need to compete with the giant public cloud provider, kind of the AWS software layer or something like that. So they are looking to change the VM to the OpenStack and then they need to create strong points or advantage compared with them. But recently I hear that some interesting case, it's a university. The university, in the university, the research institute is in the research institute. There are a lot of students, right? The students want to create the VM for some research, but the students fail that research and they need to change the topic and then they will delete the VM and then create a new VM or something like that. So the professor cannot calculate the budget and then resource for them. So they cannot use the VM or something like that. So the professor want to change to the OpenStack, I heard it, it was interesting thing, so. Well, we are from Japan, so the market is the same, but I think there are three major cases in Japan. It's like the web service like the e-commerce and they're providing websites to the customers and the second case is cloud provider and the third case is the private cloud for the development of software and testing the software. For the cloud providers, maybe the same reason for the cost effective to against for the AWS. Maybe the OpenStack is very important fact and the for web service providers they are really, I mean, they have several, my customer, our customer has scale to 5,000 host that cannot be afford for with VMware. And I forgot to say one another topic and then in Japan, telecommunications carrier, kind of the NTT or something like that is focusing on to use the OpenStack and then actually NTT docomo announced to use the OpenStack for their new carrier NFV and then they will start the services the next match or something like that. So all of the industry is watching the OpenStack. And just to add a little bit more comment in our journey on OpenStack, I think what we're seeing is specific maturity in terms of the use case that are writing on top of the basic IS layer with clear APIs. And the other point is that from a region perspective, I think what you will see because we work with all these partners is it's a different variation of maturity. But where people are really mature in markets, I would say for example, in the United States, certainly the enterprise, I mean myself and we look at data every six months, in the last six months to a year, enterprises definitely come up, e-commerce of course, media but then the use cases on top of that are very, very specific, right? For example, in Japan, what we see is that IS for service providers still a big, big concern but in the United States, for example, you will see use cases on top of it. Can you run a pass on top of your OpenStack IS layer? Your OpenStack IS layer. So I think you'll see different levels of maturity but the convergence definitely is that it's moving on the upward trend, okay? Any other questions here? Let's go ahead. What do we do about support? Yeah. Yeah. Okay. Yeah, what we did within our company, we have a good relationship with several distros like Red Hat, also HP with Helion, and also Ubuntu. And what we did was we built, well, let's say containers for several industries. If you, for instance, work with Canonical, you have the BootStack, you can do a quick implementation of OpenStack. There you have, for instance, you can use a Plumgate for the network part. And so for all different businesses, we build several containers and the support is exactly the same. So that's what we are doing. And in our case, it's a little different because in Latin America, Canonical is just starting to go to the market, Red Hat is not here, HP is not here. And sometimes what we see is that we cannot trust on Canonical or Red Hat or HP. Basically, because they are in UK maybe speaking English and we speak Spanish and Portuguese and Latin America. And people want people on site or very close, the same culture, the same language. So we do support. Basically, we install OpenStack, deploy OpenStack on the customer and we sell like a service subscription where we support OpenStack and also all the tools that we develop for the user. So we are the main contact and point of contact for the user. In our model, we can do it a little more custom. So what we offer as kind of our starting point is we'll handle the tier one and tier two for anything included in the stack that we would provide. A lot of customers want a tier one on site. They're fine with tier two being remote. They may want 24 seven. If they want that, we'll do that. If they want nine by five, we can work that out too. And then we go back to our various providers, such as Plumgrid or Canonical or Piston or any other provider that needs to be in there. And that's kind of our tier three. So if it's something that we're really stuck on, we'll escalate to tier three, but generally we wanna have the tier one and tier two as a single point, right? If something's broken on that cloud and it's made out of four or five or eight different pieces and somebody really wanted to have one specific piece, they don't wanna have four or five different phone numbers to call. They call us, we get it triaged, we do whatever we can to help them. If it's really broken, then we refer it on to the proper tier three, take it back down to them, make sure everything's fixed, and then tier one and tier two follow up, right? Make sure it's stayed fixed. And speaking a little bit on the Japan side of the house, I think that's a very good question, literally. And I think it wasn't becoming too relevant last year because we were in the POC phase. Everybody wanted to do a POC, right? So yeah, let's do POC and call 15 people. But I think what you're seeing is and the reason purposely we had these people on the stage is that I don't think so, the distributions can do it all. I don't think so the network virtualization can do it all. I think what we can do is provide with clear defined APIs a very, very simplified way of integrating different pieces together. And I think the use case will come from integrators and some sophisticated users themselves. But I think what you have to see is a transition here. Some of these resellers or system integrators have been traditionally selling hardware. That was the business model, right? Cisco, Juniper, I used to work for Cisco and Dell. That's what they did. But I think what you're seeing is a change and all of these business models are coming up because they're realizing that the real value that they can provide on top of all these integrations is support. So I think that's a relevant question and that's why I think you're seeing a shift with, you know, with Nabilio with a different model, right? Onyx with a different module. And Fairbanks with different N4. Us in Japan in particular, you know, I'm just taking us as an example, right? So we integrate with different distributions. But the reason why we went to Japan with our resellers is because they are the ones who speak the language. They're the ones who will interface. And I think as a community we have to understand that, right? It's not about 15 components and I will support it. It's not going to work that way. Business models are going to evolve. And VMware and all these ecosystem thrice because they have a very, very healthy system integrated community out there. So I think that trend and change is happening. I think it's a good question. What we did, as I said already, we are the community ambassador for OpenStack in the Benelux. We decided to use OpenStack.nl to make it very locally, as what you said, because that's very important. Also the Dutch customers, we all can speak English, but it's very good for them to have a contact nearby when there is a problem. And also we do the support for our customers. It depends on all the different products and containers we create. But also for us it's very important that the customer has a contact nearby to speak Dutch and very good. Question? All right, that would be a question for us. So the question was how well is the community doing the integrations and testing, right? I think it's getting better. And let me answer you. So from us or any other network virtualization, there's an example. There are the companies in storage doing similar things in storage. For us, we rely or work with an integration or orchestration system. In our case, the go-to market is OpenStack. And when you look at OpenStack, there's an open source version of OpenStack. But I think most of the enterprise that we see or service providers are going with distributions. So now we rely on the distributions to have a framework in which we can plug in. Looks pretty simple. You've got Neutron upstream, and that's it. Well, you realize that everybody has their own installer framework that you had to integrate with. And if you look down the value chain, what we want to do is we really do not want to get into the services business. We want to do these integration and pass it on to the value chain, right? I think, to answer your question, it's getting much better, much better. And I think I heard something from the keynote session, the way where they're evolving the core OpenStack. I think it's going to benefit everyone. But that's an area that I think still needs to be improved. But on the other side, the pieces who are plugging into these orchestration or distributions have to do their job. For example, we run pipelines throughout 24 by 7, because if some distribution does some kernel modifications, the whole stack breaks. So we are taking that ownership because we don't want these people to be spending hours and hours in front of the customers burning resources and not be able to get the stack up and running. But I think these are just transitionary times and they'll sort it out himself. Please. Testing, full of stack. Yeah, a normal week for me is I personally deploy OpenStack two times a day. I'm testing this, testing that, benchmarking. I've got something new from a partner like Plumgrid, and I need to play it against what I'm proposing to a customer. And I have a bit of a reputation for destructive testing, right? I wanted to break for me. I agree. And he will get a lot of email for me if it does break, right? So basically, yes, we don't rely on a distribution. We are distribution agnostic, but our tools are tested with all the open source distribution. So for us, it's very easy to understand. It's changing and we can change it in our tools to develop OpenStack. So basically, we don't rely on a distribution. We have a tool that connects to OpenStack and downloads the packages and deploy OpenStack in a really production manner and do it very, very fast. Okay, I think we're running out of time. I'm sure you have a lot of questions, but I would just finish up by saying that we talk a lot about technology, right? And we have multiple sessions. OpenStack adoption is also a business model discussion, right? So a lot of people don't talk about that, but I think that is what the other close source systems get really well. So I think it's about time we start having that dialogue and this was a humble attempt to get that conversation going. I'm sure there are many other ideas, but I really appreciate all your time. Hope this was valuable to you and we'll be here to take any further questions. All right, thank you, bye-bye.