Cloud World Congress into the next step, the cloud technology elite.
Tian Yun Technology Vice President Reitao: Today is really all star flashing, see the valley of the elite gathered here, so we want to pull back to China, pull back to Beijing's also Zhuang, to see what happened in the local enterprises. We invite four guests to chemistry today. First, Si Jie system company Cloud Platform Product Department CTO Liang, Mr. Liang in SI Jie company responsible for cloud product strategic planning and the overall technology implementation, once was the founder of Cloud.com, this year to think Jay. The second is Chen Shangyi, director of the company's technical committee, adjunct professor at Beihang University, postgraduate tutor of the Academy of Sciences, and a major project of the National College of Singapore and many nuclear shishou. The third is Ali Cloud computing platform, senior expert Lin early morning, Mr. Lin began to enter the Aliyun group in 2008, has been engaged in large-scale cloud distributed computing and most of the architecture work.
Finally, Beijing friends Tianyu System Technology Co., Ltd. Yao Hongyu CEO. Today are all cloud elites here, all need bright swords, to give some challenging questions to the elite, hope that local innovative content, can surpass the Silicon Valley enterprises, because it industry, heard a lot of content is speaking, success can be replicated, this sentence seems to talk a lot in China. Every successful Silicon Valley business philosophy, technological innovation, can be found in the domestic samples. More than 10 years of domestic IT industry, the Internet industry, the continuous introduction of content innovation. The elite here, not replicas, are independent and innovative print. These questions, want one one to ask us all of you here.
The theme of today's topic, I was thinking, talk about innovation, where is the secret of innovation? Ask our Mr. Liang first, there is a story behind you. Cloud.com, a team of 40 people in two years, sold 200 million dollars, so quickly in the cloud computing field. Innovation is a certainty. When we look at the Gartner maturity curve, cloud computing has not yet come to Peak Valley, still in the gestation period, 2009, Gartner or in the hot holding period, how did you do it?
Si Jie system Company Cloud Platform Product Department CTO Liang: Today is very happy to discuss cloud computing and cloud computing in China. As Mr. Ray said, I created a company called Cloud.com three years ago, our company was 6 months ago by the company bought. I now think of Jerry as chief technology officer, responsible for cloud computing technology.
In fact, speaking of innovation, the most important thing for small companies is to have a chance and opportunity, and then have such a concept to grasp, there is such confidence to grasp such opportunities. In a word, cloud computing is now very red, but two or three years ago, Amazon's Ec2,s service has been very successful. It was not just Amazon's success, but many of the industry's giants began, and Chang and Zhang were openly talking about the opportunities for cloud computing.
Therefore, I think, the most crucial to innovation, to seize the opportunity, is to start to do, do not care about these big companies are also talking about the same thing. Because from the enterprise's point of view, there are many existing products, there are existing strategies, it is difficult to quickly turn around, to do a new, now cloud base dozens of enterprises to redo, from the strategy, implementation, customer perspective can be done again. It, the traditional cloud server to the cloud, the traditional PC to cloud computing, such a great turnaround time, to the small companies like us, as well as many of the cloud-cloud.com companies here to create a lot of opportunities.
After our company is done, the most important thing to say is not technology, the key is to develop some technology to help users really solve some problems, because these problems already exist, but the company can not quickly do technology and products, strategic transformation, and really help users solve problems.
Therefore, our company cloud.com at that time in such circumstances, quickly solve these problems. At that time from Korea Telecom, India's Tata Telecom, and so on, all of a sudden there are dozens of cloud, are built with cloud.com software, we will soon be able to get these customers, and can use our software, really let customers benefit.
The next step is that big companies like some big companies, such as Jie, are also forward-looking and advocate the need for innovation. Big companies may not start, but after a year or two, to three, and realize that there is a need to invest in cloud computing and find opportunities in technology, they will acquire, in their view, the best, or the most customer, the best, the best of the small companies. This time, is in this kind of prospect, our company is bought by Si Jie company, I was very happy at that time.
Because Citrix is a fairly large company, but in this scale of companies, few companies like the company's forward-looking and ability to really want to solve the problem of cloud computing for customers, when Si Jie after we acquired, our business really pushed to the next step, so that our technology and products more successful, Let our customers more effective.
Tian Yun Technology Vice President Reitao: We need a quick choice, from an early age. We didn't answer a question head-on, how did we choose? For the next round of technical questions.
Below also want to talk about Baidu, I contacted Baidu for the first time should be ten years ago, that time, innovation, he handed me a card, is a very small green card, and all other brand card is not the same. We see the Silicon Valley has Google search company, Google developed so many years down, they are also rigorous, they also acquired some video applications, to get new cash flow. I would like to ask Baidu's Mr. Chen, Baidu decade of Development, the driving force for continuous innovation and the root of where? And talk about future cloud computing related areas of thinking.
Chen Shangyi: Yesterday evening 10:30 when there is a Weibo netizen to sleep, very uneasy, that is, Baidu a bit, he asked a question, the world has more than I handsome people? The result of Baidu's answer is sorry did not find, this netizen said, can rest assured sleep. It was a joke. But from this joke, we can feel, we can easily smile, Baidu people feel great pressure. Not only this problem, a lot of problems when you do not know the answer, will be submitted to Baidu, so just asked the host, Baidu's innovative power where? This beginning, the search engine is very simple, it is through keyword matching, find a bunch of links, and then sorted by algorithm, let you choose from, row first, certainly is said more in line with your requirements, should say.
Later even in Baidu internal, caused a little controversy, but the debate quickly passed, that is, users do not choose keywords, it is difficult to find the answer, so we need to educate users, teach them to use keywords, and soon have been defeated. Search engine gets more and more dependent, many problems will be submitted to it, such as 1+1 equals a few, such as a certain word and Chinese how to read, for example, three car together is what word, or why men than women, or to say, the consequences of cerebral hemorrhage surgery is how? Many problems. Therefore, Baidu in the ongoing challenges of these issues in the process, forcing its various technologies to continuously innovate and develop. Including the development of cloud computing we're talking about today.
Why do you say that? First of all, let's see in the face of more and more users of the problem, intelligent, fool, the search engine as a natural person, this case, August 18, 2008, in the Baidu World Congress, Baidu founder Robin Li put forward the concept of box calculation, which is a landmark, before the efforts have been. What is the box calculation? Box calculation is opposite to cloud computing, even some experts think it is contradictory, or put forward to a nondescript. But in fact, it is a very advanced technology concept.
The problem it solves is how to continuously meet the growing needs of users. Through our internal survey found that users are not only interested in data, links, now found that 30% is interested in application for the king of the era, he wants to search applications. So the idea of the box calculation proposed in 2008 is backed by four important technical support.
The first technique is semantic analysis, a description of natural language, and a small box.
Second, behavior analysis, when the weather forecast, in Beijing is to show the weather forecast in Beijing, Shanghai or Hong Kong, is to show the weather forecast in Shanghai or Hong Kong, according to the location of people, behavior habits, given the answer is not the same, is personalized calculation.
Third, the massive data, is today everybody discusses the big data, not only is massive, the relational data structure and the dynamics. These four technologies support our growing customer needs and needs. Baidu inside has a slogan, box calculation, box is the purpose, the cloud is the means. Just now, to meet the needs of users, we have to do a lot of things, four technology. But in the background to face more and more users of data and needs, unimaginable, now mobile terminals, intelligent terminals more and more popular, bandwidth is wider, upload a data is very convenient, the number of backstage is the growth of geometry, we now do is for 500 PB, or 1000 PB to prepare, There are also large-scale machine clusters.
So, just now the host said, Baidu's innovative power is what? That's it. 2008 Proposed box calculation, now after three years of development, exit the data open platform, the application of open platform, is actually a cloud support, data open platform is what, the application of open platform is what, may not have time to talk about. In simple terms, box calculation proposed, that is, that is, to search for things immediately can be obtained, do not need to link one to find. After a period of development, that is, the search is used, we can experience, is to search a small game, then the screen can show the game screen, you can play directly on the above, the current Baidu open platform application has reached more than 15,000.
Tian Yun Technology Vice President Reitao: Mr. Chen is very professional, soon Liang Jian, I understand that your answer is that the power of innovation comes from the demand, the rapid capture of the user from all aspects of the search box needs, and then found our technology of the original power, that is, application box calculation, you speak so much, semantic analysis, behavioral analysis, Machine learning, we should invite you to our large data forum in the afternoon, will focus on the implementation of the rationale. Let's take a look at one of the following questions talking about cloud computing is more, it should be Amazon, the model of resource leasing, a book-selling enterprise to build a huge data center, resource packaging and leasing, there is a new industry model, Amazon is also changing, he now also has a lot of new business to appear. In addition to selling resources, but also selling capacity, computing capacity of outsourcing.
Last week to participate in the operator's forum, when listening to a brilliant explanation, Aliyun in the domestic content, what is relatively innovative technology?
Ali Cloud computing platform senior expert Lin Dawn: Our goal in ten years, from the group's point of view, hope to create a share of data, the first platform, so we think, if we want to in this society to various companies, the industry analysis of these data, Cloud computing may be the only capability that can be provided as a basis for technology. So that's why we set up a company that specializes in cloud computing, and when we did these things three years ago, we thought about it.
One of the big elements, strategic decisions, is that we may not be using open source software like now, and we're basically starting with every line of code, and we've made a whole set of systems. is completely a few companies in the country, is starting from the first line of code, the entire distributed environment, the entire distributed operating system from scratch.
There is a very important reason for this, the application of demand is the real driving force for this technology to go forward. This is the first day is the conviction of things. We think that only in our China environment, do e-commerce environment, encounter different challenges, finally polished out of the system, naturally unlike others, this is three years, do this thing when the earliest idea.
And then back to today, we've been doing this for three years, and it's been a tough process. For more than three years, we have found that this idea has been validated in practice. We have a very small example, as we all know, the most important basis for cloud computing is that it must be massive. This is the core problem in solving the cloud computing problem. We look at large-scale, solve the better is the use of cloud computing technology companies, may be Google, a full set of GFS, Map reduce, is a big scare technology, must be low-cost, is built on a commercial PC server, rather than using hardware to solve this problem. This capacity supports the steady growth of the entire industry, which is also a foundation.
In addition to these two, and Amazon, we see cloud computing has a third feature, in addition to large-scale, low-cost, the third feature is that must be able to service, operational. In other words, it should be open, open is to make everyone together in the cloud computing platform to do things, to use, finally to achieve the goal of sharing data. This is the three points we see.
Today we look back, these three things, really do down, and then combine the characteristics of China, we will find a lot of technological innovation in it. Google does gfs, the first day, the file system is made can be added to write, Hadoop open source did not do this thing, after several versions, plus this function, because he thought there is a need.
The first day we made the file system, is to do the best to write on it, and later found that the Chinese market is not so, we know that Amazon has two very important products, one is EC2, one is S (transliteration), is that the safety factor is very reliable data in S (transliteration), EC2 does not guarantee that data is completely secure. Want to be combined to use. But the Chinese industry, the small web site does not want to use this, is the hope that EC2 inside automatically contains S (transliteration), for use is very transparent. For this, to change the background of the distributed system, the file system can not only be added to write, but also can be changed to write, in the GFS and Amazon architecture, the public architecture inside we do not hear them to do this thing. But in China's environment, the demand for such products can be acceptable to everyone.
With this demand, is really the driving force of technological innovation.
Reitao, vice president of Cloud Technology: Many key systems, distributed computing, large data, etc., also refers to the commercial and standardized low-level building equipment. There is also a key star, from the camp of distributed computing, one of more than 10 years of people who wrote Hadoop, after returning home in the past few years, Dr. Jean Yao to share with us, just heard Ali, for distributed computing has done a lot of improved innovation, Dr Yao is very understanding of the problems of Hadoop, but also made the corresponding innovation, Hadoop is open, Dr. Yao is now a privatized company operating, in innovation, there will be some collisions, such as now many do open source, open source mode, many people do not appreciate, pure open source, such a business model is sustainable? Pure private enterprise, we will lose a lot, our open source community of technical contributions, long-term technological innovation from where? Dr. Yao, while we are doing technology, because you are a business person, what is the consideration between the two?
Beijing Friends Tianyu System Technology Co., Ltd. CEO Yao Hongyu: First, from the Hadoop point of view, the original design, mainly for Yahoo to take the cloud computing platform, not open to the outside, Hadoop is not Yahoo's cloud computing platform, the earliest design for the search and mailbox storage, until now, The entire design framework has not changed much, which is why there are limitations. I believe that there will be many new innovations when Ali combines the cloud platform with the E-commerce platform. Because we are similar to them, basically put it all down, start from scratch, the reason may not be the same, and finally wrote for several years.
Back to the moderator's question, private and open source, I came back to start a business, carefully think about this question, in the end is open source to do this thing, or all from the beginning. Have half a year, is in the market experiment, open source in that period, that point, that is it can not support things, and do not meet the needs of users too much, there is no way, is all give up from the beginning.
Second, I am a staunch and user of the open source community. Previously responsible for a lot of Yahoo platform-level system, we use a lot of open source inside, but there is a point, open source really do very good project is not very much, objectively speaking. For example, the database, basically about two hands on the open source is doing a good product all belong to the. Open source projects, hundreds of thousands of of the world, should not be exaggerated.
Generally speaking, open source, many people are from the angle of interest to do this thing, the first push faster, solve some problems, but the consideration is not very comprehensive, because not as a commercial company, to serve the end customer. Open source is the original driving force is to create open source people's interest. And so to a certain extent, I think, the most appropriate way, is the power of business added to the open source of things really to the public.
So I'm a user of open source and a supporter of open source, but I'm not doing an open-source company myself. I personally think that from a business standpoint, I think it's a bit more likely to be successful, and that's the first. Second, a lot of enterprises after the success, is the beam total cloud.com part is not open source, is now being bought by Si Jie, has a business model, but this part of open source, I think, this is a very good model.
So, it should be open source plus private, friends in the short term may be a closed source system, because the market competition is very fierce, it is likely that all products will be contributed to the whole community, that is, friends can survive, can live better when.
Tian Yun Technology Vice President Reitao: There is no such possibility, we have seen many North American enterprises, in their own growth process, but also to do the whole industry chain of breeding and nurturing, very small companies are doing. Jie now uses this model in the healthy operation, also hoped has more core, like Ali, friends, these have the true knowledge connotation enterprise, can make a contribution to the industrial chain by consulting (transliteration) way.
The second question, when I was communicating with the guests, we were talking about a problem, cloud computing, big data, or the corresponding concept, where cloud computing comes from and where it goes, and we're all discussing a problem, and the cloud computing model has many ways of interpreting it. The earliest clusters, up to now, have different schools and see different camps in the environment of corporate communities. Different camps, such as Oracle, EMC, and VMware, such as enterprise-class markets, V-cloud still have many good roles to play. There are also large-scale calculations, some mature products are in use, and indeed determine the actual application scenario.
On the other hand, is a distributed enterprise, built its own kingdom, Facebook has more than 800 million users, such a large group, should not be fragile, should be a more powerful computing model. Different genres, listen to the four-bit view here. I know, all of us here belong to one of the schools, can not deny the other side of the enterprise, this angle, from where, cloud computing will go to share their views.
Si Jie system Company Cloud Platform Product Department CTO Liang: This issue is very important, although now cloud computing has been recognized by everyone, we also see Google, Baidu, Alibaba and other obvious clouds, have been very successful. Your question is fundamentally about what cloud computing is all about. In particular IT business, there are some big head, different enterprises, talking about what is cloud computing, is not exactly the same. Mr. Ray said VMware, EMC, Cisco, VCE, to do cloud, buy a machine, install the software, is a cloud, what exactly is meant. I would like to say that the definition of cloud computing is not necessary to start from the beginning, we all know, how much cloud computing, how much money, or can be remembered, there are APIs, we are very clear.
Another point, we are not very clear. When Cloud.com was established, when Si Jie bought us, why did these things happen? There is a very common understanding of what is on cloud computing. At that time, many big IT companies were selling cloud products and doing well. But what they fundamentally do, in our view, is not a real cloud. In our view, the real cloud companies are network companies, such as Alibaba, Baidu, Google, Amazon, such as companies, not traditional to the IT industry, sold to large enterprises to do IT companies things. Why is the distance, the traditional or very large machine, the more the bigger, more and more expensive, look at Baidu, Google's system how to build, built a very large, but itself every machine is very cheap, and the beginning of the program and application, take into account the following infrastructure, may be a single look, not very reliable, Each individual may be thrown away, overall, because a lot, not only in a place there, all over the country have, Baidu, Google services like this, I have not encountered problems, maybe the internal machine in the bad, we can not see.
Traditional it is difficult to manage, to be in charge of Cisco, VMware, EMC, all need to go to class, all want to certify, just like take a doctorate degree, the knowledge is very big. The new cloud computing content, thousands of tens of thousands of machines, the biggest is think Jay with our products to do the private cloud, do nearly 20,000 physical machines, not virtual machines. So, every day in the running, every day the machine is dead, die, is to change a new machine, the application has not been affected by a little.
I think, from this perspective, what is the cloud, what is not the cloud. Jie's point of view, this really can be like network companies to build data and it is the real cloud to build. The traditional it to get, write a little thing, everyone with how much statistics down, and then do User Service, to do a terminal, is not a real cloud. Therefore, our company's goal is to find ways to learn Google, Baidu, Yahoo, Amazon and so did what, many large enterprises do not have such a technical force, do not come out, the big network company to do things, we make software products, do a good job, the oil companies can use, hospital departments, government departments can be, Most are open source software products, so that our customers can do the real cloud infrastructure, but also in this way.
Tian Yun Technology Vice President Reitao: Thank you! The view is very clear, the gunpowder is also very sufficient, there is no manufacturer on stage today, but fortunately our afternoon eight manufacturers VP, will have the right to speak, there are opportunities for resistance. Baidu's Mr. Chen Unscramble.
Chen Shangyi: Liang always said the situation of Baidu more comprehensive and accurate, I agree. If you have to ask where and where to go, I think these two camps, internet companies certainly pioneer, and then the second camp, another camp, they do is product and service. The infrastructure, data processing products, why do you say so? Everyone knows the fact that cloud computing should start with Amazon and Google, like these internet companies. Why did they bring it up first? The last link I talked about, they have a very challenging and unprecedented demand, Baidu's data and Baidu's machine cluster, the size of the data center, is one of the largest in the country.
So, in such a big case, who provides the technology to put all the distributed cluster management up? So big data, who's going to teach you? Maybe Google, Amazon have, but will never give you, who depends on? by our own power. I'm talking about what you know, and there's a lot of things you don't know. I would like to cite a few examples, such as timeliness, we received your questions, within 0.05 seconds of rapid response, users do not want to, will choose another browser. Advertising, take money in Baidu advertising, eager to come out immediately advertising effect, here is how large calculation, advertising and search results of the correlation between the calculation, in order to display ads in the right position. Also hit a word, immediately appear a hint, hint too late is no use. Just say now the hottest time hit a word, can be prompted. All of this is forcing you to do a lot of computing behind you, and to achieve very high real-time requirements.
I mean, internet companies are solving problems that other people haven't solved.
Tian Yun Technology Vice President Reitao: I understand, the Internet companies in the Pioneer, in the first attempt, and then these big enterprises, but also explain why, these large enterprises are talking about Hadoop, all need Hadoop story, we do first, finished, there will be large enterprises packaged, and then to the enterprise.
Chen Shangyi: Yes, the traditional companies do not need to worry, the Internet is to do the service for the vast number of users, never thought of making products, sold to a third party, even if there is a subjective desire, there is no such ability, this is the core competitiveness of the Internet, not willing to expose to others. So give the second camp a good chance.
Tian Yun Technology Vice President Reitao: From Ali's point of view? Ali's point of view, we are the first practice, or we have other rigorous way?
Ali Cloud computing platform senior expert Lin early morning: The first two speak better, I add a bit of my own opinion, today, cloud computing, not to define what is cloud computing, is an ecological circle, different companies play different roles. We want to go back to the cloud, mentioned earlier, why the Internet companies will be in front, we look at the beginning of our electricity, the beginning of electricity, is a big factory first used, electricity, the real human society occurs when the most important benefit, must be to the people electricity. This perspective, cloud computing which day really become a public cloud, become everyone, every small business can use, this is our future cloud computing where to go.
Then again, we have always understood why we define the goals of the group's next decade as the first platform for sharing data. A very important question is that we want to share the data, which is equivalent to the way electricity is used, to greatly promote the benefit of the whole society, to solve many problems in the society, because the data chain has been cut, resulting in the problem of not optimization, this is our idea. So, from Ali's point of view, what we want to do in the future is the service of the public cloud, which is certain. Many people began to have doubts, we set up a cloud company, is not only for Taobao, not, we provide public cloud services. There will be third-party software companies, solution companies, for different industries, in the public cloud services, packaging private solutions, vertically provided to the real use of their businesses. This may be the whole biosphere of the future.
In addition, including individual developers, they in the cloud era, in fact should be able to enjoy, more innovative environment, which is also very important. Because ultimately this thing, cloud computing is very exciting, really with the electricity came to human society, the real excitement is not after the electricity came, is thousands of electrical appliances out, enjoy the role of electrical appliances. Therefore, we think of the different cloud applications are produced, greatly generate social benefits, is the real wealth of cloud computing.
Tian Yun Technology Vice President Reitao: I understand, Ali still rely on these technologies, provide the appropriate service to the final party. Friends, seems to be between the two, friends with these technologies have walked into the enterprise, Service enterprise users, Dr. Yao have any ideas?
Beijing Friends Tianyu System Technology Co., Ltd. CEO Yao Hongyu: I will not tell you friends, we are to listen to technology. I've been involved in the whole internet for a long time and I'm talking about the ins and outs of cloud computing. In China, I started to do cloud computing and promote cloud computing in 2007. Now there are customers who ask me whether I am doing cloud computing, I am confused, is now the market is flooded with too many fake cloud computing experts. I agree with the previous experts that cloud computing was born from internet companies. Whether it's business model or technology itself. The earliest time to be born was the late 90, is to wait for the internet bubble More and more, the user volume, data more and more, the system is more and more complex, no other ready-made methods, no one has thought how to solve this problem, so will be born, formerly called large-scale distributed computing, 2006, someone took the cloud name , everyone likes it, it becomes cloud computing. Large-scale distributed computing, now with some partners, or the super large-scale distributed computing, technical perspective.
From the point of view of service, now whether it is Ali want to do service, Baidu want to do service, Amazon want to do services, are services. From my point of view, is a spiral of internet services, the Internet originally provided content services, to the current Internet 2.0. Now on the Internet, you see all the publicity is IaaS, SAAS, PAAS, many people still foggy, is simple cloud computing, service perspective, is the Internet 2.0, the original poor service provider now as a service to provide out, technically speaking, is the extension of the calculation, now the name sounds good, called cloud computing, these two together, will produce huge productivity, for society.
So, I agree with Ali's comrades that the ultimate dream of cloud computing is that the public cloud is used by all, without the so-called private cloud. But it's a very long time to get to that point.
Now this stage, is the private cloud or more prevalent, even in the future dedicated aspects, even if everyone can use electricity, there is a large enterprise or their own engine.
Second, the cloud computing genre, the earliest time, cloud computing, everyone did not mention, may be the company's decline, the first is Yahoo, Yahoo's cloud computing platform, the internet has something, they all have. He made a cloud computing platform that was born to support a wide variety of applications, and that's what the cloud should have today. Cloud computing platform perspective, Yahoo's platform is the most like, really we are now mentioning the cloud, but it is not open.
The most famous cloud computing platform, currently running the biggest cloud is Google. Google is typical of its own business after doing a good job to open their own business, very appreciative of them. Hadoop open source products from Yahoo, but is not to take the idea of Yahoo, but Google's thinking, go to the GFS mentality.
The two camps that the experts mentioned before, traditional manufacturers and the Internet, I do not think so. Technically speaking, the Internet domain is divided into many branches. Google's most fundamental technical point of view, based on storage, storage based on all the applications, VMware, Citrix from the virtualization perspective, extending other applications. IBM, Oracle, their technology accumulation is very deep, they are now in the original market, now need to let others to share, but also need to change themselves, they feel very uncomfortable. In their own traditional products, the introduction of a lot of new elements, is accelerated slowly modified, which is also a genre. From our point of view, we are a little sibuxiang, friends do, I am equivalent to experience the whole process of technological development, from the design of this thing, from the problem-solving, and finally to the operation of large systems, to discover the cloud computing platform, the core problem is how to make a lot of machines, whether virtual nodes, physical nodes, or storage resources , and let them work together to accomplish one or more things. Work together, with people, institutions and institutions to manage the same, Synergy is the fundamental communication, there is synergy based on the basis of synergy, storage, etc. are based on this, so it is not the same as other aspects.
Tian Yun Technology Vice President Reitao: Thank you! I have heard, in fact, today's four guests are from the perspective of innovation, challenge the traditional giants, Oracle, IBM and so on. The last time, Hadoop iteration problem, friends how to solve? The closeness of the data is poor, how do we solve this problem?
Beijing Friends Tianyu System Technology Co., Ltd. CEO Yao Hongyu: I do not know what the source of this problem is, compared with what, because a lot of things are a comparison, to speak worse. First, the goal of Hadoop's first design was to search and store the mailbox, without considering it. Hadoop is now used more and more widely in structured data processing. There is also the problem of iterations, not too much of a relationship with Hadoop, even if there is a scheduler in the database to do iterations. Or it would be better for Hadoop to do the iteration itself, or it would be to dispatch a resource outside. Data relevance is difficult to solve, the database to do performance improvement of one of the most important technical considerations, is to put the relevant data together, visit, you can read it at one breath. In distributed systems, because most data technologies are implemented by using hash (transliteration) methods, there is no correlation factor. Sit very close, hash later, maybe one in the United States, one in China. The hash algorithm itself can be adjusted, that is, if the real Data association, do not put the hash itself in the traditional concept of technology, data relevance can be done.
Tian Yun Technology Vice President Reitao: A lot of differences, can extend a variety of applications, thanks to four guests, have the opportunity to listen to the manufacturer, the VP from eight manufacturers, see how they explore cloud computing and indigenous innovation of the elite of the duel. Thank you! This is the end of today's forum.
(Responsible editor: Duqing first)