 I hope you're not too tired. So that presentation on the business track is about launching a cloud service provider with OpenStack. So my name is Philippe Theriot, and I have my colleague, Raphael Ferreira, in support of me if needed. Let me build a background around me. I've been with Inavance for the last few months. Started with the cloud computing business around 2000, when we actually have launched one of the first cloud service provider in Canada. At the time, the word cloud was not invented, so we were calling that application service provider, but delivering the same benefit. I've been involved in the launch and creation of many cloud service providers on various platforms on VMware, CloudStack, OpenStack. And in the last few years, I was an advisor to investment firms, mutual funds, investment companies who were investing in many cloud computing businesses. And I've joined Inavance last June as the chief commercial officer for international expansion. Just quick add on what Inavance does. Basically, we do three things. The first thing we do is that we design, build, run public cloud infrastructures for telcos and service providers, all four countries, especially following July 6 for those countries who want, let's say, non-Western-built cloud computing environment, or just without the NSA API. So this is what we do. We also do exactly the same thing for private cloud, mostly for enterprises, for the ones who want to re-internalize certain workloads, and really for the ones who want to launch a private cloud that is inspired by the public cloud services, similar. And then we've also managed some specific applications on top of multi-clouds, what we call that multi-cloud management. Basically, balancing and distributing some of our customers' critical application on top of Rackspace, Amazon, Google Engine, HP Cloud, and some others. We're at Office Base in Paris. I'm more on the US side with Office in Montreal and San Francisco, with our office in Bangalore. And in the next couple of days, an office in Singapore, an office in Singapore will be open and also an office in the Middle East. We were 50 employees when I joined last June, and now we are more than 100 employees. So that presentation is about launching successful cloud service providers. And of course, the main topic is about the conditions to launch a successful cloud service provider, knowing that probably no one wants to fail in that business. And the conditions are quite simple, but also quite complicated. Actually, the recipe is simple, so you just need, let's say, good ingredients so that you can consider that the technology. Then you also need the recipe. You can improvise if you want. But the best thing is to follow a recipe, and then you need a chef. So you choose. You either go with this one or that one. And since we're on the business track, that could be on technology, but unfortunately, for the geek in you, to successfully launch a business, it's much more about business than about technology. That's the way you will look at things. That's the way how you will justify the technology, and how you will finance it more than the technology itself. And it's all about that. For those of you who studied in marketing, that's the holy grail of marketing. That's the holy grail of successfully launch something. You need a good product. You need a good go-to-market strategy, sales, marketing. Of course, you need a good technology. You need a good foundation for your business. But if you fail in operation, your business will be a disaster. If you fail, if you have the best product in the world, but you cannot sell it, also your business will not be a good success. And all of that is nothing today. If you're not the first in the market, the market shrinked, although it's a market that evolves every day. So you need to add your own value. And for us in IT that sometimes that's the most complex things to do, adding value, identifying what's our value. You can look at the market saying that the market is full, and there's so many providers that it's totally useless entering that market. Or you can look at it and say, yeah, you have something unique. You have your own value added, and then you can compete in that market. So since we are OpenStack Summit, yes, I'm going to talk about OpenStack. And actually, OpenStack is an amazingly good engine to start and to leverage to start your own Cloud Service provider. Whether it's an internal Cloud Service provider, if you've reached that level of maturity, or if you want to become a public Cloud Service provider. However, OpenStack, and that's interesting, is also a very interesting, very powerful business enabler. And just based on the conversation we can have here, that's a good example. So back to the diagram, basically, your obsession to be successful in that business has to be your added value. And then that leads to one single thing, a famous three-letter acronym, your unique selling proposition. So how will you differentiate, or your service will differentiate from others, from the other services available in the market? So that's the ingredients that you need in your recipe to have a success. So now the route to success. How will you reach the success? So basically, to find the route, basically, you need to find the problem. And sometimes finding the problem is kind of tricky a bit. And for those who have a solution for problems that basically don't have a solution, that's the best way to make a lot of money in that, actually, in general. Either you create a problem, or there's a lot of companies in the IT industry that creates problem, and they solve patches to solve the problem. But otherwise, you have to find a problem. And sometimes, especially the technical guys inside of us will look at a certain situation and say, OK, there's a problem here. I'll give you an example. So you could consider that, oh, there's an opportunity. People want an online CRM. Is this a problem to solve? Not at all. Actually, what people want? So people are tired of upgrades, patches, and complicated CRM data replication synchronization. That's a problem. So then, if you go to the market with an online CRM where the data are ever synced, the way you will tailor your marketing, the way you will organize, the way you will go to market will be totally different. If you focus on basically the pain, the upgrade, the patches, the synchronization, that's the way you're going to sell your product instead of if you focus on the fact that it's online. Because you will basically spend a lot of money trying to convince people to go online instead of basically having them imagine a thing, how it's going to be easy when their data will be ever synced or where they will not have to manage patches or upgrades. So here are a few problems that basically you can solve. And actually, that's a mix of, that's a hodgepodge of problems and kind of ideas. But commonly, where there are opportunities for public cloud providers, it's where the resources are hard to find, where it's complicated or it's too expensive to find skilled resources. A good example now around the Bay Area, finding, with basically San Francisco San Jose area, finding a system administrator under $150,000 a year of salary, it's barely mission impossible. So basically, where in the rest of the United States you can find system administrator for probably half of the price. So that creates a problem that is killing some economy. So if you have a solution to that, so basically you can make money. So the lack of support for ever changing IT environment in certain areas. Too many vendors involve doing finger pointing, especially in the small and medium enterprise market, where I'm always used the same analogy, who will they call? When they're, they will not be able to attach a Dropbox document to a Gmail email on their iPhone. They need support. And if you have a solution to that, you're going to make money. Support or interface in specific language, data hosted in specific geography, or the need for provider compliant with certain regulation. In the United States, we have, let's say, the HIPAA for health care information patient protection. That's SPIPIDA in Canada. That's another acronym in France. That's another acronym in the UK. So basically, service providers are looking for, that's the SAS providers in the health care business, are looking for specific providers, yes providers, or past providers that have those certification. So basically, they're looking for something. There's a problem. There's a need. Then you need to find your value added. So what will you add to the mix? And that mix of a problem and your value added will be your unique selling proposition. So basically, that could be geographic proximity. That could be low latency. You're close enough, or you are in certain geography. I know that in Mexico, there's not that many infrastructure as a service provider, although there's great needs. You have an end-to-end SLA, especially in the telco where we see a lot of demand. An end-to-end SLA is something that could be only delivered sometimes by telcos because they can take control of the SLA at the router and basically deliver the whole stack. You could do billing in local currency, and you can do billing locally. It's very hard in Venezuela. They have access to a certain pack of US dollars. So if you can bill them in Venezuela in their currency, actually it's much easier for them that if they have to go to the bank to basically pay you. You have fixed price, where everyone has variable price. You have a vertical solution for a specific industry or a specific sector. Let's say, what about a community cloud for a pack system in the health care industry? If you build your own network, you build your own data center, or maybe you partner with someone, there's a lot of demand for that. What about a community cloud for order kind of, it's in the media and entertainment business? So you have specific skills. So for media and entertainment, that's a good example where if you have specific skills in the streaming it, let's say for streaming servers, then you can build something specific. You can offload a burden of the shoulders of a lot of, let's say gaming or media and entertainment businesses. So sources, so your energy must go on that. Sorry to upset you, not necessarily on open stack. Your energy must go on that. Otherwise, you'll have the best technology in the world, but you will never be able to sell it. So is there a market for you? I have interesting discussions with a lot of big vendors, a lot of big telcos, a lot of dominant players. So that's a garden magic quadrant. Everyone recognizes that. So those are the big guys. That's a crowded market. So should you go in that business? Is it interesting in entering that business and competing against those guys? Actually, you can take the problem with many angles. The first one is that they're too big. And actually, you cannot compete. You will never be a successful public cloud service provider. And if you want to tackle and if you want to basically enter the same market with exactly the same value proposition or unique selling point. However, you can say that they can't solve any problem. They'll never be in every geography. They will never be close enough to every customer. They will never have all the skills that your customers will want. And I believe that we only have the early stage of migration, from a lot of servers and storage volumes, from on-premise to off-premise. So basically, there's a lot of need that will have to be fulfilled. So the right way to look at the market from my perspective is that those big cloud providers don't have and will not have 100% of the market. So who has a Dropbox account? And actually, who has more than one online personal storage account? A lot of you. So who has only Gmail? Who has only Outlook.com or Outmail? So even in our personal life, we have many providers. There are such things, such as, let's say, Setupbox, the satellite TV that's kind of a hard to have two providers. Actually, even cell phones now is a big thing. I'm just amazed at the number of cell phones with dual SIMs. So cloud are consumable products or solution. And people in businesses can have more than one supplier. It's exactly like a fleet. Like a fleet for any enterprise. They will have pickup trucks from Hyundai or GM. And they will have, probably, I don't know, like sedan from motor vendors. And every car basically serves a purpose. So the multi-sourcing, which is basically from many external providers, but also a super big trend that is basically the multi-sourcing, the mix of internal and external resources, is really, really becoming a hot trend in the enterprise sector and also in the small and medium businesses sector. I've worked for very large corporations where basically we had six different CRMs serving different purposes. So what are the platform or the framework options? And I didn't try to boil the ocean here where it's just few names to identify the family. There's more than 40 vendors in that space. Some are bigger than others. And actually, just behind Cloud Stack or just behind Open Stack, there's many vendors, actually. So the IBM way of things, where basically IBM and some other equipment vendors developed large cloud computing platform, and they have a huge ecosystem of resellers. What they're trying to do, they're trying to convince their resellers to basically, instead of reselling physical server, to resell basically cloud servers. Greater margin, actually easier to attract the customer, easier to attach the customer. And IBM has been very clever in the past with their choice of, let's say, dropping the personal computer business. Now, actually, they dropped the small server business to enter the cloud market. So it's a trend. However, is it a good choice for you, if you want to enter that market, to have a dependency on that? To basically have your customers calling IBM instead of calling you, having IBM invoicing your customers instead of you. So who will basically, how easy will it be for you to attach your value at it if you go in that direction? So we see very new business built on top of that. Then we have the Windows Azure. So you can go speak with Microsoft. I had the opportunity of being on an advisor board for service providers at Microsoft for many years. And it's funny, because they compete against their ecosystem. So yes, if you want to become a cloud public provider, you can go and you can sit with Microsoft and you can launch your thing. However, they will compete against you, actually. And they will never send you business. So the only difference that you will be able to have is basically local proximity. Yes, Microsoft will never have data centers in every jurisdiction. I think that Myanmar might be a good place to have a data center if you are in that jurisdiction. So Microsoft will not be there. Cuba, North Korea. But at the end of the day, it's servers and data could be installed and hosted everywhere. And yes, you can attach your value added, but you will always have the fear that one day someone can come with exactly the same thing. VMware goes exactly the same direction, although it's a very good solution. It works. They have a very mature ecosystem. That's a solid, robust solution. And it's easy to have a lot of value added. There's a lot of, I think it's blue luck as an example that has a lot of value added on data center, on replication and backup and recovery services. So that's a good platform to start and build a business on top of it. However, they will compete against you also because they also sell virtual machines on the internet. Then you have the pack of commercial solutions. Once again, I just put three names here. You have the UNAP, the Parallels and the Joyant and many others. Those are very good solutions. If you want to go to the market fast, it works. It's proven. It works really well. And actually, most of them, they are fully integrated or they have pre-can agreement with, let's say, Zoroa or Ubersmith for billing and invoicing. It works. Or you can go on the open source. So you have, I guess everyone knows, the Eucalyptus CloudStack and OpenStack. I'll not enter in the debate. It's better to go with CloudStack, OpenStack or Eucalyptus. I used to be a strong believer of CloudStack. Actually, I've launched service provider with CloudStack. And now, I think that OpenStack is mature enough and has a much more brilliant future than the order platform here. And it's easy to see where on the slide, where you have all the dominant players in the Magic Quadrant from Gartner, 75% of those players are open-stack or running an open-stack cloud or are migrating to an open-stack cloud. I think that's a bold statement. So it comes to basically a few things. So your decision criteria. So what's the best way to add your value added to solve the problems? So what's the best platform that basically will support you and will give you the means and the ways to enable, to bring your value added to your customers. And then that you will differentiate yourself amongst the others. Because your unique value proposition has to be unique. And then what makes your service competitive, profitable with a price structure that will resonate? And when I've mentioned that, let's say those, there's a lot of, you have a lot of choice. You can pick and choose anyone. Actually, that is running. That is a closed ecosystem. However, probably when you will look at the numbers, those license costs or those integration costs or maintenance fee will just kill your business case. So you can have the best solution in the world if you don't know the price that your customer can afford or if you pay too much, your business case will be a disaster and your product will never actually leave the ground. Please invest in a matcher and complete ecosystem to interact with. No IT is a complex world. No one has the monopoly of good solution. When I started as a consultant in the 90s, I was selling, I was working for a large three letters consultation firm and I was selling like a body shop like super experts in security. Those guys, they became an expert in anti-virus and though actually those guys, they became expert in intrusion prevention and actually finally I've realized that the same guy is now a super expert in anti-spam. You need more and more precise expertise. So the bigger the ecosystem around you is, the easier you will find and you will have access to resource to support you. That's why I'm not a big fan of those commercial solutions because basically they try to nail you down in a sales process where basically all the doors will be closed and you will have only one number on Speed Dial, which is their number. Ease of finding technical resources. So that's why I became a fan of OpenStack because now the communities matcher, there's a lot of people on OpenStack you can find, just see how many people are here. So finding technical resources is a challenge. In certain geography it's even more complex so make sure that you pick and choose the right platform that will support that and basically where you will have access also to go to market expertise. People that you can call and that basically did exactly the same thing that you are trying to do just in different market or in different geography. And then an integrated or complete solution to let you work on your core competency. There's a lot of promoters that will try to convince you to reinvent the wheel. Remember I said you have to invest, to be successful you have to invest in your core competency. So reinventing the wheel is certainly not a good approach to be successful in that thing. So basically by finding a platform where you could have something that is kind of a package, I'm also a fan of package because basically you have a package but the package sometimes has boundaries that will limit your ability of adding value added. And of course avoid vendor lock-in which supports what I was just saying earlier and compatibility with the rest of the world. The old concept of the internet, the whole concept of the cloud is exchange of information, exchange where volumes and data could move and migrate. So it has to be exactly the same in your business. So why OpenStack? So OpenStack, you have to admit that Inavance is a French company, I had office in France, although I'm not French so that's why I use a Boeing example, not an Airbus. OpenStack, sorry guys, OpenStack is a ninja where it works out of the box where you can tweak it and the way you're gonna tweak it basically will make your airplane just much better than other airplanes. That's why when Boeing decided to make the Dreamliner they just didn't take the, let's say Boeing 767 engine and just rip and replace. They basically work with GE and Rolls-Royce to develop new engine. So OpenStack is a good engine however you will need to tweak it to make sure that it's gonna work for your business case. It's gonna be stable, it's gonna be fast performing, it's gonna generate whatever you wanna generate in the time that you think your customer will want it. Then the OpenStack community will always be there to support you, basically. So OpenStack will deliver you the engine, the rest of the community and even your own skills basically will build the aircraft. Components, you know that Boeing and Airbus and Bombardier, Embraer and others are now just assemblers. So they are taking parts that are manufactured everywhere in the world and they just make it together however they work on the design and they make sure that everything is certified. So the OpenStack community is now mature enough, has a lot of help that basically can help you do that and then you will bring your value added. Every airline has access to exactly the same number, same kind of aircraft. Your value will be where? The way you're gonna do your layout, your business class layout, would you go with a 10 address or a nine address? Your route, of course in the case of Kataraways, the cost of fuel might help a bit your business case. So where you will locate, and actually that's a key business advantage that you might have. Then the service, the food you will serve, is basically the way you will add your value added and that's exactly the same analogy you can apply in the airline industry. So why OpenStack? Because that's a large community. So you're not alone. It's backed by big logos. As I said, 75% of Garner IS leaders are going to OpenStack or own OpenStack but also because you will not be locked in in that thing. You will have OpenAPI to go hybrid cloud with order vendors, with orders providers in your ecosystem and it's gonna be easy for you to add value. It's open source. You'll be able to play with it and if you play with it with the right partners, everything that you will do could become upstream and then will be accepted and actually you will never be alone. And you'll be part and you will be integrated in that ecosystem. You have the option of do it yourself or you have the option of going with an integrated solution that basically will give you the ability of launching your service just much faster and some are 100% open source, 100% upstream and there are all, however, there's also very good solution that are taking the core open source and just adding proprietary software on top of it which basically brings a bit of lock in of course. I'll speak about, let's say, our reference architecture for that, where the way we can do exactly what I just described and we're not alone in that business. There are order businesses that does that. We're basically everything that is read is OpenStack. So you bring your OpenStack, we can take the core OpenStack, the original OpenStack or some commercial distribution of OpenStack and then adding all the pieces that you will need to launch all open source to launch a public cloud service provider. It works and you can go to the market actually really, really, really fast. So those kind of solutions now exist and actually if you look at the diagram, it's kind of a bit similar to what you have on the commercial side with the Joian on app or let's say Citrix Cloud Portal. Exactly the same but now we've reached a point where that exists in OpenStack and actually this is what we do. Add in of us for private cloud or for public cloud or for public cloud inspired private cloud. So your business advantage by going open source, of course you'll be able to change some components. If you don't like the mod, if you don't like let's say jibbling, can replace jibbling. Their API actually that's an open API. So if there's someone else who has another open API or maybe the connector could be developed and if it's accepted, it will be supported. It's backed by a commercial entity where basically you are not alone. So those companies who are in that space can help you build, run and also support. So just giving you more time to work on your unique value proposition, your unique selling point. So as you work on that, basically someone else is working with you and then if you wanna basically take back, control back of your solution, then you can go. And because it's packaged, that's a proven design. It works. You can go to the market faster and in some market it's really important to go fast because every customer that you're not, there's transaction every day and actually if you are not out of the market, actually you're losing customer every day. And the beauty of that is the beauty of the cloud is recurring revenue that you can get. And actually you will have just more features and you will have a predictable budget and predictable outcomes most of the time. So that's the advantage of going with open source but also with pre-packaged open source solution. So what about the hardware? So when launching actually one of the biggest capital expenditure will be on the hardware and we're often often asked, should we go OCP or commercial mainstream hardware? And I strongly believe that if you wanna compete against Amazon you better use the same hardware. So they go what, let's say OCP style server. So if you wanna compete, if you want exactly the same, the way we address the market is that when we work with customers, they can sell at Amazon price and still make a lot of margin, more than 50% margin. So if you wanna be in that space, you better use that. If you use very expensive hardware you will probably not be able to have that same margin or even actually better because we're working on, especially with OCP vendor to even cut the cost more. So I see OCP on the volume and velocity and commercial mainstream more on the value and service. So you have to define yourself. It's very rare that someone will play in both fields. So now we see that hardware in the volume and velocity business hardware adds very little value. However, you have to pay close attention because that contributes a lot to your bottom line. That's gonna kill your business case. However, if you go on value and service hardware adds a lot of value. You probably wanna bring those important logos that's a EMC, NetApp certified, storage, whatever it is. You might have some interesting connectors to negotiate with those guys. And what's funny is that we see now a migration. We work with a lot of our hardware vendors and a good deal on some UCS or a good deal on NetApp. When it's a good deal it's sometimes you can get the best of both worlds. And those guys are interested in getting in that business because they know that when they win a customer it's for a long time. Although OCP is very mature and if you deal with the right person. So if you have the right support. So who will support? Is it an integrated solution? A bit like what we do in Inavance where we come with pre-package pods for some specific needs for volume and velocity where it's easier for us to support because we built the whole thing. However, there's some other cases that the hardware will be provided by someone else. So and can you afford to go with OCP if you go volume and value and service not necessarily. So what kind of reliability? If your customers will run cloud applications you don't care about the hardware. The redundancy will be built inside of the application. However, if your customer run Microsoft Exchange Server on top of that I would recommend let's say a traditional server. And it's not a one size fits all and you have to speak with the right guys. In those hardware vendors they have teams that are not your local teams. They are dedicated to service providers. And those guys now actually I've asked many years ago can I pay my networking ports per port sold or active and now they do that. Same with storage per terabytes sold. So if you speak with the right guys the guys who are able to serve service providers and we can introduce you to those guys so you will get a much better deal. So one last thing and that's back to the beginning. So don't forget to sell. When we look at customers, when we look at cloud service providers who are not succeeding it's because they just forgot one thing. They forgot to sell. If you want to enter that place your main focus should be on your value added but your value added if you're not able to sell it. If you don't have a powerful e-commerce website you will never be able to attract those valuable customers and get those interesting recurring revenue. And that's why actually that's the first thing you should have in mind. So how will I sell the thing? And then when we look at the vendors the lineup of vendors there's very few who have a powerful e-commerce website on top of their solution. Something that you will be able to cookify your prospects when they will visit your site. You will be able to enable some re-marketing with banners and campaigns and activate analytics on top of your site. That's the first thing to have in mind because sales is actually cash is king to finance that thing. And then you need quick results. You need to be on the market fast and then you need basically customers that you attract. So in conclusion, say stick to your value added and your unique selling proposition. That's the thing that's your uniqueness about your future business. Find a partner for the rest. Don't be alone. No one has the monopoly of good solutions but also no one has the monopoly of knowledge. OpenStack is certainly now an excellent framework with a lot of potential that can deliver great results but more importantly great profitability for many kind of businesses in that industry. There is a possibility that OpenStack is not suitable for every needs. If 100% of your customers are running VMware, I guess that you better call VMware. And don't try to reinvent the wheel. The ones who are not succeeding, it's because they try to reinvent the wheel. End user portal exists. An e-commerce website already exists. Don't develop a CRM. I ever actually run into someone yesterday that redeveloped an entire billing and invoicing system. That's not value added contributing to your bottom line or that's certainly not a key differentiator unless you have a very strange currency. And don't forget to sell your solution. So that's how you could launch a successful cloud service provider with OpenStack. Still have two minutes for questions? No, actually there's many. So UberSmith has an OpenStack API. We recommend our customers jibbling. We've developed a connector and it's 100 compatible with our OpenStack deployments and it's in production with some of our customers. And it's very, a billing solution that has to be customized for recurring revenue. That's a nightmare. If you try to do recurring, I tried doing recurring revenue in SAP. I don't recommend you do that. But those are two examples and actually we have many examples that work well with OpenStack. All right, thank you very much.