Peak conversation: Big data is a lie?

July 12 News, 2012 China CIO Summit Forum held in Jiuzhaigou, the following is the main title "Big Data is a lie?" Peak dialogue.

Sun Yu: Welcome you back, our afternoon meeting to continue, the morning of our activities, in micro-blogging business launched a lively discussion, the Jiuzhai activities we use Sohu and Tencent Weibo, users can send: #2012中国CIO高峰论坛 #+ arbitrary content participation, Mobile phone users to participate in the way please see the big screen below, welcome attention and participate in the on-site interaction. I'm senior editor Sun Yu of the IT Manager World magazine. Very happy to host the event this afternoon, Forbes magazine published a "Big data is a big lie" article, to those big data proponents poured a basin of cold water, according to the article, from the database to the data warehouse to today's big data, The rapid growth of data volume is a trend, and the corresponding data processing methods are changing and improving, which is a spiral of the development process. In short, the big data is not the end of the world, nor is it an alien, then the enterprise's IT department does not need to pose formidable posture, big data is not a lie? Today we have invited several heavyweight guests to discuss this topic together. Welcome to this discussion guests: Feng Chingsi, General Electric China's Asia Pacific IT service Director, Mr. Shenlitao, deputy general manager of Eerduosi Technology Co., Ltd., Mr. Teng, senior vice president of Customer Prudential (Beijing) Technology Co., Ltd., Mr. Lee, senior vice president of Jingdong Mall , Mr. Chen Jinxiong, Director of Information Center of Fuzhou General Hospital, Nanjing Military Region. The big data is 2012 Enterprise's most important word, I want to ask each guest you in the mind big data is what, each guest's front has the pen and the paper, asks you to write out your mind Big data. Lee always wrote a Mass, excavation, action, Teng always wrote is a mass, diversity, technology, Ms. Feng Chingsi is written by the data flow, mobility, diversity, Shen always write is the method, relativity, universal load, Chen always wrote the data volume, value, management, we ask you to explain the key words you wrote.

Chen Jinxiong: Large numbers, as the name suggests, massive data, the second massive data must be valuable, there must be the potential and value of excavation, the third should need an advanced tool for management, if there is no good tool to manage, this data will not be reflected.

Sun Yu: Shen always the more philosophical nature of your writing.

Shenlitao: For us to say methodology, after large data is raised, from the design of the data, to the production to the processing, is actually a methodological system, this system for our business operations bring a certain impact, the second it has a relative problem, a we say large data when linear nonlinearity, are some relative factors, In our application level, we are in order to adapt to our needs, how to do the relative data of the planning, the universal is very strong, in such circumstances how to get useful data out, this is a very big difficulty and problem.

Feng Chingsi: I think big data is just a noun, in the past, so many applications in it, we are always dealing with the data, the past may be small data, is now large data, we used to use the Data Warehouse, we are static data, data are stored in the inside, now a lot of feelings, is actually a stream of data, including a lot of software that we've used recently, including the encounter Sogou a navigation software, in fact, there is a data mobility and mobility, so more to my feelings is the data flow, in fact, in this data seed how to capture more useful, more practical, more business value of information is more important, This includes the fact that we look at the data from different latitudes, so I think it's more about data flow, mobility, and data diversity.

Moderator: Lee Total Teng is always doing business, your first word is massive, you talk about digging and driving it.

Lee: If we dig out something of value, if we don't use it well, use it to make decisions, use it for our open source and river closure, or use it to a greater extent, that's not going to work.

Teng: I want to say that now a difference is diversity, the reality of the data you know is not just some text data and numbers, in fact, there are more audio screen data, another with the development of technology, we now have the possibility to do this particularly large scale data analysis, For example, it is difficult to imagine that ten years ago or 15 years ago analysis of TB on T is completely impossible, but with the development of hardware, including some software, including some of the development of the structure, more is the analysis of a large number of information.

Sun Yu: Thank you for your views on the big data, please write down the big data is not a lie. Five guests all the choice is big data is not a lie.

Teng: All the vendors who sell big data solutions are all lies to them, but we as users don't feel like lies, so not necessarily.

Sun Yu: Below I would like to share with you the work of their respective enterprises, on the large data have some problems, is how to solve, for Feng said big data This work means what?

Feng Chingsi: In fact, the big data has been discussed for a long time, I said very frankly, in fact, there was no landing of the solution, I do not know if the company here, we have only a few practical solutions from last year, I also take this opportunity to share with you, the first we in the United States, In terms of energy we have a GE business unit, this wind has a lot of data analysis, including how to cloth the wind turbines, including wind direction, geographical environment, air detection, and geographical conditions of the data are all put together, and the wind turbine has a lot of considerations, and finally after this large data analysis, to make a , it has wind turbines within an effective distance, the first to reduce waste, and the second effective application. I'm talking about two big data cases, we have a lot of cloud computing, I used large data to do a network monitoring, network monitoring We know that we have hundreds of points in Asia, hundreds of points including the company's internal operations and radiation to various hospitals, medical this aspect, there are thousands of tens of thousands of points of network equipment collection, network system Operation Collection, This is now in the cloud, but our front network monitoring is a real-time monitoring, we set a target, he traced to the network failure, or network utilization use high will be an error, this is a real-time data monitoring, we also have set the value, if this network of abnormal alarm, there will be a system to do deep analysis , is it because of the instantaneous control problem or the instant network problem or indeed network problems, through this we can detect the network health, we used to do with static database analysis, but we now have these dynamic tools, we can provide to the business unit, or hospital, We can make a solution for this. The second is also medical, we set up a call center in China, the customer's data in the call center is also a lot of, such as the hospital called, the hospital because of the purchase of medical equipment, call to consult some questions, or reported a failure, we also set if the connection rate in 15 seconds and broken, There's a data that's broken within 15 seconds of calling the caller. The call center at the front desk also has a video and interaction, and there is an interaction between them, as well as an instant messenger and instant messaging that forms an agent to our call center and also helps the user solve the problem in the first time, So this is one of two applications in the Asian region.

Sun Yu: Feng Total mentioned two examples are medical aspects, in fact, we Chen is the Fuzhou Military District Hospital, the medical department it may be all the data types of the most complex one industry, do not know you have encountered this problem in the work?

Chen Jinxiong: I think the development of data, we study philosophy is to engage in research process, what big data? Developed to a certain extent, right, need new technology, such as a unit enterprise, its decision-making, its support all need to rely on data to support, it should be into a large data era, which is similar to our car, car may have developed for decades, more than 100 years, The original car is a rich man bought the car, but the car became our necessities, become our way of life into a car era, so I would like to first on the first question I have a brief discussion of my personal point of view. The second really feel the role of the data is very large, we have to deal with GE also very much, originally may ct,2000 years of time a CT only how much? 10M, now the CT has been 320, or even more than 600, a patient to do a full-body scan, the data volume has been to dozens of G, now in this view, how large is this data? Now our medical equipment, a lot of information collection, why? With the improvement of medical level, the improvement of medical diagnosis, the precision of the data more and more, the second from the management point of view, management of fine, as far as possible to each point of data collection, I remember in 10 I wrote an article, Digital hospital, which I mentioned a full range, the whole process, This really produces a very large data for us. There is also a hospital inside the amount of data is really very urgent, the complexity of medical treatment, influential, there are graphics, there are standards, there are non-standard, there are tests, there are medical records written by doctors, different specialist medical records are not the same, many people say that the medical industry, pharmaceutical information should be the most complex one of the current information.

Sun Yu: Internet companies may face large data is the most direct, I would like to know what the most pressing problem for customers?

Teng: A place where internet companies have a lot of advantages, for example, regardless of the customer as the online e-commerce business, it is the best relative to the traditional enterprise is it very easy to collect data, I yesterday with Wanda's friend chat, just know Wanda can through the camera to count, how many people to enter every day, how many people have come , no need for E-commerce, we are born to get almost all the data on the site, this is a very large amount of data, it is also a very large protection, how do we analyze it, take Fank as an example, we collect data every day about hundreds of G to a T order of magnitude, How do we analyze it in such a large quantity? This is really the challenge and the problem that we are facing here, the head here, like just mentioned, it has a variety of data, it has pictures, it has voice, it has this data, how do we find what we need here, of course, we are relatively young guests, we are also starting from up to do some large data research, For example, the distributed approach, structure, and even some new ways to do some research based on user behavior, but we are still at a very early stage.

Sun Yu: That want to ask Lee the total problem is jingdong every year, the most concern is actually the opening speed of the Web page, and the processing speed of orders, this is the most concerned about, Jingdong encountered large data problem?

Lee: The problem we have, in fact, in the first quarter of this year, it was more than the amount of data we had in the last three years, on the one hand we think that valuable data, we have to record to save more data, on the other hand, the growth of the company's users, this is also a greater pressure on us, on the other hand, the pressure to deal with, You find that the traditional solutions are out of the way and you have to adopt a distributed solution that is a big challenge for us.

Sun Yu: Want to ask Shen Total, because Thai is to help enterprises to build and manage the data center, in your work, from the point of view of the need to meet customers what large data needs?

Shenlitao: Well, I've actually met a similar demand, in fact, the main problem is that we analyze from the customer's point of view, mainly to feel his data on the density of the value of the problem, so I think these four problems inside, I feel a part of the customer inside, very typical, is because the value of large data is relatively low, The amount of data that is invalid is very large, and as a customer it is not possible to determine what is valuable, even if you can determine, there is no good technology can be extracted, this is the biggest problem, bring a few problems, there is a waste, if this data can only save three months, may be the fourth month of the time before the deleted, It's a waste of data, the second is to constantly expand his storage, our own storage will never keep up with the growth of data, this is the problem we now face, of course, our project hope that through the introduction of cloud computing technology, through the IT architecture of the redesign and design, to help it solve the problem, But I think now is not a fundamental solution, I think the next step in any case, we as a Chinese enterprise, must be from the data keyword, from the relevance of the design, as soon as possible to put this up, this solution may face the less difficulties.

Sun Yu: Just sharing the big data in the process of the problem encountered, the next question is what you think Chinese enterprises in the big data applications of the gap and bottlenecks are? Also write three keywords. Director Chen you have written, to show you what is written, ideas, technology and applications. What do you think is the most important explanation?

Chen Jinxiong: Remember we held a new Digital Application challenge in 2010, I said three, concept innovation, technology innovation, application innovation, in fact, the concept of the key, our data you admit not to admit it is there, then how to do? We want to have new ideas, business or normal management, to know how to use data for our decision-making, management, process application or provide support, I think the idea is the first important, the second technology, that is so large amount of data how to manage, just said, A big difference between this big data and the general data is where it has to be large to a certain extent, just Jingdong Lee always said, big to a certain technology will produce new technology, the original can be resolved, now to the distributed solution, large to a certain degree of need for new technology to solve this problem, so technology, the third is the application, Regardless of the idea, ultimately for our application.

Sun Yu: Shen always tell you your words.

Shenlitao: The first foundation, the second requirement, and the third implementation. The foundation of this piece, for our large data does not exist such a problem, for some of US state-owned enterprises have this problem, none of us to talk about the big data, because we have to do the infrastructure, the second said the demand, I worked abroad for several years, then let me spend a week to learn his report, I had to do some research first, this is the cultural differences between East and west, in China, I think the next demand is the same, because we now data structure is from the west, which we have a cultural differences, so the future in the demand of how convergence, this is our next challenge. The third is the implementation level, the implementation level we are speaking from two angles, the first we from the current database solution is very high, are open source architecture, these architectures if done two analysis, the requirements of people is very high, to some enterprises can do, but there are many enterprises are unable to do, There is a high demand for our practitioners here, and another at the infrastructure level, we now want to big, high level of data centers to deal with large data, although I am doing data, I hope that we use more data center, especially in the unclear time, we make small, flexible, flexible deployment to operate, May be better for the business.

Sun Yu: What's your key word?

Feng Chingsi: I think the first one is the infrastructure, the second is (tools), the third is visualization, if you mention China's bottleneck I think China has a few bottlenecks, the first China's existing compared with other Asian countries, if the maturity, availability and scalability is only higher than Thailand and Vietnam, India, far less than Japan, South Korea, Australia, because I am in charge of all over Asia, so I have a more detailed data for the entire structure of Asia, China in this regard I think there is a bit behind, even if it is compared with Hong Kong, Taiwan, may still have a certain distance. The second one is on the data center, our data center, which I used to use in Shanghai two months ago, is also a telecommunications ctidc, and I think there are a lot of infrastructure issues on the big data, it has storage, security issues, and if our infrastructure isn't there, it's a little too early for us to talk about big data, So I think China's infrastructure is a bottleneck. Maybe I'm a bit one-sided, just my personal point of view. What are the second tools to implement these? because you, in addition to such as medical imaging, image transmission, you may have a share of the platform, the real data is useful to you, this is a bottleneck, the third is from my CEO that level, we have such a chart, data in the back of the analysis, Really can be rendered in front of the visualization of the graph, this visualization I think is a problem, the example that I said just now is that the network bandwidth has done a large data research, we appear to be able to understand the figure, the people in this room can understand it, not necessarily not the IT people can read, So the third visualization, I think, is a bottleneck and a place to develop now.

Sun Yu: Lee always tell you?

Lee: I wrote three words, the first one to pay attention to the degree, because the data it has value, but you do not cherish it, you do not think it is a treasure, it is not enough attention, we may be the internet industry, because it is basically driven by technology, but many of our traditional industries, our CIO status is not high, The importance of technology is not enough, the importance of the data is not enough, so the importance of the degree is the key. The third is to drive, to use, if you still shoot the head of the management method of this decision method, rather than technology-driven, data-driven management mode, in fact, this large data you are not good, I can say two examples, an example is our web page, you may go to see when you will feel very simple, But each page is made up of a lot of blocks, and there are different strategies for each block, different strategy output is different, its conversion rate, we use conversion rate measurement, so any one enterprise can use different points to test various strategies, choose the best strategy, this completely by the data to drive our web page interface, This is a scene, another scene is like the BoE every month has a management analysis meeting, it will analyze a lot of data, actually just talked about visualization, visualization may be on some reports, some statistics is more appropriate, but in fact, we have to do fine management, we often have to do the analysis of the topic, These analyses may have an idea that we are validated by the data, and we innovate to change our approach and finally update the results to improve the level of our refined production.

Sun Yu: In Teng always show the board before I tell you, Teng total began to write down the talent, innovation, and finally waited seven seconds to write down the last one you say is what?

Teng: Attention, must pay attention to, attention is not enough, I write talent innovation, we now use a lot of distributed and so on storage way, in fact, this aspect of the talent from our now look at the relatively small, or relatively scarce, and on the other hand from the innovation, in fact, many are American technology, Few original, domestic now a lot of reference to the so-called micro-innovation, I think this is very good formulation, but the real big development we need very structural significant innovation.

Sun Yu: Teng has just said that micro-innovation may bring hope for our country's innovation, our last question, is please write down what you think the future enterprises in the large data operations to achieve what breakthrough, write one to three keywords. Teng always write the fastest, the application of this too big, you explain, in the end is the industry application or what application?

Teng: As I just wrote, you collect so much data, if not be valued, cannot make a contribution to the decision, it is nothing, it must be applied to really can play a role.

Sun Yu: Lee is always you write open source, throttling, promotion.

Lee: Income and expenditure can be analyzed from two aspects, a business like ours is two things, one thing how you reduce the cost of your customers, he has a lot of channels, he constantly measured, choose those good channel for him, so you will find that he is relying on data analysis to promote, in fact, this is a kind of open source, The other is how to maximize the value of orthopedics, orthopedic came, you put it in the bag of money out, then to analyze what kind of person he is, to compare the behavior of other users, and then to optimize his conversion rate, we have a comparison, with this personalized recommendation technology, and without this recommendation technology, they will be 30 less sales % to 40% so that's a typical example of an open source, I just said refinement management, through data analysis to improve efficiency, which is very typical throttling, reduce costs, and just said the promotion, in fact, we in this process, we can through data-driven approach to improve our management, To elevate the level of our use of data, like the one we talked about today, which is a cyclic process.

Sun Yu: Lee always has doubts about whether big data is a lie, but there is no doubt that using big data to get more money from customers ' pockets. How about you, Feng?

Feng Chingsi: Input, application and quantification, I think it is the first chicken or egg problem, there is no input, there is no output, of course, after the application can be applied, for example, we now talk about a lot, in fact, cloud computing is a large data business model, the application, how to bring to the enterprise what benefits or benefits, Or bring him some sales improvement, this is what we think of the IT people here, otherwise we always say to ask for money, how good, how to analyze, after analysis of what, for example, including user behavior model analysis, after the analysis of the two out, to you this user growth has any influence, this is to consider the problem , so big data I think is a core of cloud computing, or a soul bar.

Sun Yu: Shen, your answer is?

Shenlitao: There is a breakthrough in data outsourcing, why I say breakthrough, we now many companies are doing similar to the application of the work, if let a company immediately take a team to do data analysis, it is not an external force implementation is very difficult, I believe that in this area, China will be a breakthrough in the next phase, And I believe that the Chinese people's wisdom, for this kind of mathematics, to this kind of data understanding degree, I believe this work will be very good, this is also i a very good anticipation, we hope to see more and better data outsourcing companies, in China early appearance, thank you.

Sun Yu: Director Chen your answer to show us?

Chen Jinxiong: The first focus of attention, an in-depth excavation, the third canonical data, I was the first to speak, and finally summed up the speech, because our data is for use, from how to use this data, large data I talk about point of view, one is the focus, micro-blog, now Sina Weibo has 3, 400 million of the amount of daily release, forwarding, Reply may have billions of of the amount, through the attention of people, the topic of concern, it can. Morning I told Sun editor said, I said we must run the idea of journalism, the idea of running a newspaper to pay attention to the data, the news is so much, is also a huge amount of data, the newspaper's front-page headline is what thing? Must be the reader's greatest concern. Attention to this data is not to say that other data does not look, by focusing on one of the biggest bright spots, we have key indicators, through a most concerned information, and then through the deep excavation to find it, you see Weibo has attention, he has search, through in-depth excavation can find the data you need, third data must be standardized, So the normative data is the basis, so we talk about marriage, what is Chinese marriage? Married, unmarried, you see our international distinction, there are premarital cohabitation, a suspect marriage, married, have a divorce, and then live together after the divorce, he this marriage situation, your personal life situation, on your disease, or the situation of your life, on your physiological and psychological effects have a change, so structured, standardized, standardized, and standardization must have the theme of standardization, I was in the hospital, so I proposed that the hospital must take the patient as the center, the clinical as the center, to the doctor as the center, the patient to see a doctor, a medical treatment is a clinical center, how do we use medicine? How to do the examination, how to do the operation, all of these should be around the patient this main line, I will briefly express this view.

Sun Yu: Thank you today for your views of the collision, let us come to the big data is not necessarily a lie conclusion, good, thank you! Please sit down.

(Responsible editor: The good of the Legacy)