 Thank you to all of you for being here on a weekend to talk about open ecosystems. I'm very impressed with the number of people here and the fact that such good questions and such good dialogue. And so many tweets. I don't know how many people are following postings, but we have a real tweet storm going. Hundreds of tweets and retweets about this conference. Let me show you my Twitter ID. If you want to contact me, that's the best way. I'm Mike Nelson on Twitter. If you're still using email, which is so 20th century, you can reach me at mnelson at pobox.com. So those are the best ways to reach me. If there's anything I say that you want to follow up on, any references I give you that you need more details on, please feel free to contact me. I am now a professor at Georgetown. I'm a visiting professor of internet studies. Georgetown has never had one of those before, so I get to be the first one. And as a professor, I give lots of reading assignments. So through this talk, I will give you at least four or five things to read if you want more information. I'm very grateful to be here again. This is my second visit in six months. I'm just incredibly excited about what's going on in Qatar, and I'm very excited about this meeting. I think this meeting has the right people talking about the right issues at the right time. Not only is this open access week, I think that Qatar is on the edge of doing something really important with creative commons, with open source software, and with the open IT ecosystems that I want to talk to today. As my bio says in the program, I've been working on the internet for 23 years, trying to shape it, promote it, and make sure that it serves the most people in the most ways. Like Joey, I look a lot younger than I am. But I have to say, in all those 23 years, we have never been at a more exciting critical point in the development of the internet. And I've lived through the commercialization of the internet, the explosive growth of the web in the mid-90s, the development of e-commerce, and now the development of social media. The next two or three years are incredibly important because three very powerful technological trends are coming together. First, we have the development of broadband wireless so we can be connected anywhere, anytime. Second, we have the development of the internet of things. In five years, we could go from having two billion people using their PCs and smartphones, connect to the net, to a world in which there are one trillion different devices and sensors somehow interacting with the internet. Most of those will be very cheap and inexpensive RFID tags and sensors, but that's a big jump from a couple billion to a trillion. A world where we have 50, maybe 100, maybe a thousand devices that we interact with all wired. So that's the other big change, but the most important change is the development of the cloud. And the cloud works synergistically with the other two technologies to unleash a lot of new opportunities and completely change the way we are doing computing. But it will only work if we build a cloud around open standards, open source software, and an open culture. And that's what I'm going to talk about today. Let me start by defining what the cloud is. In the United States, we have the National Institute of Standards and Technology. They've been working on standards for more than 100 years. They spent the last two years trying to figure out what the cloud is. And they brought together different players, the technologists, the users. And they really worked through this question of how to define the cloud. So let me read to you in two sentences what the definition is. And this is the first assignment to you. If you want to understand the cloud, look at the full definition. They have a one-page paper that describes what this cloud is all about. But the main definition is the cloud is a model for enabling convenient on-demand network access to a shared pool of configurable computing resources, e.g. networks, servers, storage, applications, and services, that can be rapidly provisioned with minimal management effort or server-provider interaction. What that really means is that computing becomes something that is provided on-demand self-service mode. You can access it anywhere. And you can pay for what you get. That changes the whole game of computing. It means that you can get all the processing power you want, a lot more affordably than you can today. It means you can buy as much storage as you need. It means you can scale instantly. So you can go from serving 1,000 customers to serving 100,000 customers in just a few hours. You can easily integrate different applications because it's all running in the cloud. And you have lower maintenance costs and lower training costs because you're using a standardized platform. I can go into a lot of detail about how the cloud works and I urge you all to read the NIST paper. But I will describe the three different flavors of cloud computing. The first flavor is sort of fundamental. It's called infrastructure as a service. And that allows you to just buy computers on the net. You bring all the software. You provide the operating system. You provide the applications. You program it to do what you want. That's infrastructure as a service. If you want more, you can buy platform as a service. This is what Microsoft sells with Windows Live. There are similar services that provide you with an operating system upon which you build applications. And then the highest level of the cloud is what's often called software as a service or applications as a service. S-A-A-S. And that's what most of us experience when we use the cloud. Facebook is an example of that. Flickr. The various backup services that you use to back up your hard drive. All those are examples of applications as a service. The bottom line with all this is that within five years 80% of all the world's computing and all the storage could be done in the cloud. I say could because while the economics and the technology are pushing us in that direction, we do have some barriers. But we're moving fast. I think we're moving even faster than we moved with the web back in the mid-90s. And I would argue that the cloud is even more important than the web and will be even more disruptive. It's going to change the winners and the losers in the IT system, in the IT arena. It's certainly going to empower a lot of new companies. Companies which today don't have the skills and the resources to build elaborate websites and powerful new applications will be able to do that because they'll have access to the cloud. There are two other pieces of reading I'd recommend. If you want to understand how companies are using the cloud, particularly small businesses, I'd point you to a survey that was done by The Economist magazine two years ago in October 2008. It's 14 pages, about seven articles, and it's entitled Let IT Rise. And if you just do a Google search for Let IT Rise, you will find this series of articles, an excellent overview about what the cloud is, why it matters, and how it's being used. Those of you who want to do some additional reading, there's a great historical book called The Big Switch by Nicholas Carr, and he compares the cloud and the shift from individual systems to cloud computing to what happened 100 years ago with electricity. And he goes through and looks at the economics and the changes in the culture that happened then and compares it to what's happening now. It's a very nice book. It's not technical and very well written. I mentioned earlier that we have some challenges. First off, we have to change people's mindsets. We have to convince CIOs and CEOs to migrate off of their own systems onto these third-party services. We have to take care of security and privacy concerns. A big piece of that is ensuring we have identity done right. We have lots of different ways to authenticate people online to establish identity. We need to be able to do that for the users and we have to do that for the providers of all this equipment that makes the cloud work. We need standards to make everything work together. And for some applications, we have to know where data is moving in the cloud because in some countries there are restrictions on what data can go where. We need audibility. We need reliability. One really big issue is we have to build systems that ensure that users are not locked in to one particular vendor or one particular technology. And the last issue, but not the least at all, is education and training. This is a new way of doing computing. So computer science departments have to change what they teach. We need testbeds that those students can work on. Otherwise, we'll slow down the adoption of cloud technology. We solved all these issues for the Internet itself. Although we're still working on security, we've made a lot of progress there in the 15, 20 years since we've seen the commercial web. But we're going to have to move even faster with the cloud. For those techies in the audience, I would recommend another paper. It's entitled Cloud Computing Use Cases. If you just do a search on cloud computing use cases, you will find an exhaustive 40-page report on all the different ways in which cloud can be used, all the different applications that are being developed. In order to address all the challenges I just listed, we have one really big problem that has to be addressed. This overarching issue has to be addressed properly, or else we're not going to get the rest of this right. We're not going to get security. We're not going to get interoperability. We're not going to get systems that avoid lock-in. And that is simply this. We have to make sure that we have a truly open, interoperable cloud. The other option, of course, is ending up with several different companies running proprietary systems, each running their own cloud service, which don't work together. I've written a paper called Building an Open Cloud, where I describe three different scenarios. One of them is called the mini-clouds scenario. That's the nightmare scenario, where Google's doing one thing, Amazon's doing something else, Microsoft is doing something else, and you can't move your data between these different systems. They don't work together, and they use their standards, their own proprietary technology to lock you in and to lock up your data and try to dominate the market. That's kind of what's happened in the U.S. with cable television. And the other end of the spectrum is what I call the open cloud, or blue skies scenario. This is where we might have thousands of different organizations, all running different pieces of the cloud, all running on common standards with common interfaces, so that we have a cloud of clouds all working seamlessly together, just as the network itself is a network of networks working seamlessly together. That's the dream. In between we have a sort of hybrid compromise, you might call it the hazy skies scenario, where different companies are running different clouds, it's rather difficult to move your data back and forth, but at least there are some interfaces, there are some ways to move data around. But it's not like having a unified global cloud. One model you can point to of the open cloud is something called the open science grid. How many people have heard of the open science grid? Go home and read more about it. This is an incredibly exciting project. It started in the U.S., but it's spreading around the world. Every day they provide a million hours of computing cycles to over a million separate research jobs. Most of them very tiny, but many of them add up to very big projects, very computationally intensive challenges that are running on computers scattered in dozens of locations. Any given job might run on a dozen different sites all tied together into this one open cloud running on open standards. Two days ago I was at an all-day meeting on the cloud in Washington, and one of these speakers was David P. Anderson. He is the head of something called BOINC, the Berkeley Open Infrastructure for Networked Computing. They are probably the most extreme example of a cloud of clouds. They do all their work using volunteer computing. Most of the computers in their cloud are simply PCs. They have hundreds of thousands of PCs which have extra capacity, which isn't being used. Those systems are all contributing to the BOINC cloud, and they're doing research on genetics, on cosmology, on diseases, literally hundreds of different projects. Really exciting example of what can be done, and they're doing it for almost nothing. They can provide 10 teraflops of computing, which is the equivalent to one of the largest supercomputers in the world. They can provide that for about $100,000 a year. To build something that big would take about $1.5 million to rent that many cycles of computing from Amazon would cost about $4 million. So for 2 or 3 percent the cost of a commercial service, they're getting almost free computing done. So what do we need to do to get this open cloud to build the blue skies scenario? Well, we need to force openness. We've done it before. Back in the mid-80s, there were many different networking protocols. Most of them were proprietary systems. You had Apple Talk, you had DeckNet, you had SNA from IBM. Each one was controlled by a different company built around proprietary standards, but TCPI came out of the academic community and because it was open, because it was open source, because everybody could build to it, it became the dominant mode and it brought together all these different networks. We did it again in the mid-90s. At that point the web had been defined and now various companies were selling web browsers and several of those companies decided that they were going to improve on the web. They were going to add their own proprietary add-ons. They were going to give you a better experience because they were going to have special things built into their browser. If we had gone down that path, we would have ended up with websites that only worked well with particular browsers. You would have really seen a bifurcation or a trifurcation of the web. But again, users stood up and said, no, we don't want that. We want one web, we want standards, we want all browsers to work with all sites and that's the way we went. We have to do the same thing again with the cloud. It's going to be harder now. Back in the mid-90s, there were maybe tens of millions of dollars, tens of a few billion dollars at stake. Now we're talking about all of computing, or at least 80% of all computing. There are trillions of dollars at stake over the next 10 years. A lot of companies jockeying for position and a lot of them trying to build their own proprietary systems, satisfy their monopolistic tendencies. We have to push back against that. We also have to prevent patent abuse. Some companies are trying to claim key points, trying to block other players from building on their cloud technology. We have to worry about nationalism. There are countries that are saying that they have to have their own cloud. Their data has to stay in their own country. They don't want to use one of these commercial clouds that are being provided by companies in other countries. We have to worry about censorship and about control. Government's trying to control where data is stored, how it's done. Many governments don't like the idea of thousands of different pieces of the cloud run by thousands of different organizations. That's harder to control. Much better to have one or two companies that they can go to. We have to worry about the national security agencies in certain countries wanting to have access to your data, my data, that's stored in the cloud. In our country right now, there's a major debate on how the rules about search and seizure and wiretapping will apply to the cloud. I know if I store data on my own hard drive and I keep it in my home office, that the police are going to need a search warrant to get that data. If I take that exact same data and I back it up in the cloud, it's not clear how protected that data is and whether a search warrant is needed. So there's a group right now, the Center for Democracy and Technology that's trying to work through those issues. Another really big concern of mine ties into what we heard earlier about Creative Commons. I'm very concerned that Hollywood and content creators are going to look at the cloud and think it's just one giant digital video recorder. They're going to think, oh my God, all this copyrighted material is going to be stored in the cloud and they won't be able to charge you for it. That's not why the cloud was designed. It won't be the dominant use of the cloud. But because some people could use it that way, there are people pushing right now for technological fixes, legal changes that would make it very hard to build the cloud as we envision it, particularly one that's scattered in many different locations, because they want to know that nobody's going to back up copyrighted material to the cloud without permission. Another very big concern is security. We all worry about security online. There have been some crazy proposals to change the way we do the internet and the way we do the cloud that could break this concept before it ever gets started. And we have to worry about overreaction to security threats, things that would lock down the cloud in a way that would make it almost unusable. So those are our challenges. What can Qatar do? Well, first off, Qatar, like all governments, can buy smart. Qatar governments, one of the largest purchasers of IT equipment and IT services in this country. That's the case in most countries. Governments have billions of dollars that they spend in this area and more and more of them are moving to the cloud. So buy smart. Buy systems that are built around open standards. When you buy cloud services, make sure that the company can give you a migration path. If you decide after two years that that cloud company is not serving your needs, you have to know that you're going to be able to move your data and your software without a lot of effort onto a different cloud service platform. Second, it's very important to fund open research grids. I mentioned earlier that one of our challenges is to make sure we have students who are learning how to use the cloud. By building a test bed, by building something like the open science grid, you'll be able to train the next generation of computer scientists for this new platform. Third thing to do is to be a pioneer in using the cloud in the government. The cloud is one of the top four priorities of the White House Chief Information Officer, Vivek Kundra. And he's been pushing all the agencies to move onto the cloud, save millions of dollars, and provide better service, better security, and more opportunities for innovation. And last, Qatar could be a model. Could be a model by developing policies that are good for the cloud. These range from adopting intelligent rules on online piracy, making clear what the rules will be regarding wiretapping in the cloud and how privacy laws will apply. Those four things, buying smart, funding research, being a pioneer in use of cloud in the government, and adopting policies that will foster the cloud are ways in which this country can really make a difference. The reason I'm here is because I believe you can really make a difference. You're a million people, but you can have an impact around the world. I've written a paper that you might find useful. It was done for the OECD in Paris last September. They published a briefing paper for a conference on cloud computing and public policy. You can either find it on my website at Georgetown or just type in cloud computing and public policy, and you'll come across this. It provides sort of step-by-step what needs to be done if we are going to ensure that old, outdated rules don't stand in the way of the cloud. In the old days, in the old world, a country had power if it had lots of land or it had lots of people and soldiers, or if it had precious commodities, oil, rare earths, whatever. In the new world, soft power matters too. Perhaps more. Qatar with Al Jazeera has shown the power of media. Countries like Denmark and Estonia and Singapore have shown the power of a good model. They've shown how IT can be used in e-government and how you can really take advantage of that to provide for a more efficient, more productive, more innovative country. But I think perhaps the biggest way to have soft power in this new world isn't through media or models. It's through magic. Magic technology. What's the most important thing that Estonia has done in the last 20 years to change the world? Skype. Magic technology. Completely changed the business model for most international telecom companies. Qatar could provide the seed for a magic cloud. A cloud that could be used for all types of innovation. Research, media, collaborative communities online. You could make a huge difference by putting in place the right policies, the right platform and bringing together the right community at the right time. And now is the right time. Thank you very much. Happy to take questions. How much time do I have for questions? Okay, two minutes for questions. Okay, I'm sorry to take so long. I do respond to tweets, and I've been tweeting all morning, so Mike Nelson from Twitter. So when you talked about storing things in the cloud, now you also have to think about where in the world is it stored because if it's companies like you said in the United States, they export all the data, not in the cloud in the United States, but maybe somewhere where privacy laws would help protect it. I mean, how do you oversee to overcome that while still keeping things in the more open sense? Well, a key part of building the open cloud paradoxically is going to be making sure it's a private open cloud. You need to have auditability so that we can show the customers of the cloud where their data has been, who has accessed it. The power of the cloud is that your data is scattered in many different places. You're not going to lose all your data because one server goes down. Your data may be stored in three different places, two different countries. But we have to have that auditability. We have to have clear rules and transparency. I have to know how my data is being treated. I have to know what rules the company is living by. And that has to be sorted out. I don't think we need international treaties for that, but we do need clear business contracts, just as we have clear business contracts today between an ISP and its customers. I'm going to take one last question real quickly. Right here, yeah, right there. Yes, I want to talk about the cloud from two visions, that making cloud or using cloud. Is there a new idea to make Cattery have cloud servicing for Warlet to make servers in Cattery to serve services, cloud services, open cloud services to all the other Warlet? Or we are talking about using another open cloud services from other countries? That's a great question, and the answer, of course, is both. Many of the companies here are going to be using cloud services that are commercially provided by countries elsewhere. But I think Qatar could also be investing in its own piece of the cloud and joining operations like the open science grid or some of the other research grids so that your own universities will have access to these platforms. Your students will be able to learn how to use them. But it's a great question. I'm glad you gave me a chance to clarify. And I'm sorry to have to cut it off. I will be here for the rest of the day. You will find me on Twitter, and you'll find me online at mnelsonatpobox.com. Thanks again for being such a great audience, and thanks for the chance to be here.