Large data needs to end data islands

Almost everyone realizes that big data is changing people's lives, and that it will bring about a complete revolution, including technology, products, industries and even the entire economic model. A government-issued policy is just the right way to encourage people to develop big data. On the same day, the State Council issued "on the promotion of information consumption to expand domestic demand," made clear that by 2015, the scale of information consumption over 3.2 trillion yuan, an average annual growth of more than 20%, Led the relevant industries to add more than 1.2 trillion yuan, of which the new information consumption based on the internet reached 2.4 trillion yuan, an average annual growth of more than 30%.

The face of such a tempting and huge cake, whether it is a traditional IT companies, or in a variety of data for many years of the internet companies, and even telecom operators can not help but the excitement: IT companies such as Oracle, IBM sits in advanced technology, internet companies such as Baidu, Alibaba is in advertising push, Personalized marketing and other aspects of the first practice of large data technology for many years, the three operators also grasp the other enterprises can not match the real and huge data sources, the parties with their own advantages to compete in large data, have made a piece of the idea.

In fact, the development of big data has long been the focus of the Ministry of Industry think-tank, the Telecommunications Economic Experts Committee. Caixin reporter learned that at the end of 2012, an internal communication, dozens of from academia, industry, government departments of experts focused on the topic of only one-big data.

In the discussion, the big data from the theory to the practice of innovation is booming, we need to pay attention to this emerging areas from where, in the end, how to develop the situation, how to widen the large data plate? To this end, the Caixin reporter interviewed the Ministry of Industry and Information Technology Institute of Telecommunications, Dr. He Baohong, director of the Internet Center, in his view, under the impetus of technology, the former humble data suddenly become a resource, but also a can create great value assets. Only the application of this asset is still in its infancy, and there is no corresponding policy to guide these applications.

Big data must be successful.

Caixin Reporter: Now the market is talking about big data, what is the definition of large data?

He Baohong: Frankly speaking, there is no clear concept for big data. In Wikipedia or other web interpretations, "data that cannot be processed by traditional tools" is referred to as large data and some attributes are added, such as "in Effective time".

In my opinion, the big data focus on how to deal with "big". "Big" means large capacity, memory, rapid change, relatively speaking, also refers to the ability or tool to deal with this data, since it is large data, it means that the analysis, processing, application of irregular and has been changing data.

Caixin Reporter: At present, the big data has developed to what extent?

He Baohong: Big Data is not an industry now, it is embedded in the cloud development, the scale is still very small. Its development is still in the initial stage, has not yet grown to be able to be independent from the cloud computing. It will take at least 35 years to separate from the cloud.

In contrast, cloud computing has passed the concept of a period, is at a rapid growth stage, and now the big data, like 35 years ago, the cloud is still in the incubation period, we are talking about exactly how to do products, how to have a market. In short, big data in the bubble stage, cloud computing has blown the bubble, pragmatic development. However, although the large data is just the beginning of the technology, but this technology to meet the needs of society, will certainly succeed.

Caixin Reporter: What is the relationship between big data and Internet of things and cloud computing?

He Baohong: The Internet of things can be seen as a collection of large data, cloud computing provides a common processing platform for large data, but it is not enough to rely on the cloud computing platform, it is necessary to do some work on the platform of cloud computing.

The relationship between the internet of things and large data is far from the same, as for cloud computing and large data, like operating systems and database management systems. Large data is based on cloud computing infrastructure services, almost every large data processing to rely on the cloud platform.

Who is the big player

Caixin Reporter: Why does the concept of big data erupt now?

He Baohong: Any technology is not for nothing. Before 2000, everyone was committed to study the traditional database, the structured data processing; After 2000, with Google, Amazon, the internet giants as representatives, began to deal with unstructured data, and use data mining results to recommend their own products or put ads.

Until 2011, 2012, after 10 years of practice, research, the internet giants finally through constant technological innovation, found a cheap and efficient way to deal with various kinds of irregular data, and from this data processing, application benefits. In a profitable situation, and this may not be a small profit, other industries are naturally willing to move the practice of Internet companies in their own industry, so that the concept of large data is packaged, that is, nearly two years of things.

Caixin Reporter: At present, relying on large data, can there be a successful profit model?

He Baohong: Now, the most successful applications of large data are some internet companies. Baidu can analyze user semantics, understand user habits, hobbies, in order to push ads, Alibaba can also use data mining to carry out accurate product marketing, based on user browsing, search and other behavior analysis of user needs and then push ads or products, is currently the most typical large data application mode.

In fact, when big data really develops, there may be unimaginable applications, just as Google can predict the epidemic by analyzing the keywords people search for, so many innovative applications will follow, and the space for development is too big for us to predict.

Caixin Reporter: Big data is now the main application in the Internet field, the specific case? Can it be used for reference by other industries?

He Baohong: Take Taobao as an example, this E-commerce platform has more than 1 billion kinds of merchandise, the total amount of transactions has exceeded the trillion, every day about 3 billion times web browsing, tens of millions of commodity transactions. So many commodity data, user data, transaction data, social data, etc., through analysis, excavation, the application of the final form to the whole process of the transaction, including the user potential purchase demand forecast, targeted push products, product satisfaction survey, business reputation, and even to the payment, insurance, logistics and other links, Derive a very imaginative application.

In the process of applying large data to Internet enterprises, some general data analysis methods, data development tools can give some reference to other industries, but specific to analyze what data, mining out what value, create new applications, but can not copy the Internet model, should be in accordance with the specific requirements of different industries, enterprises to do.

Now, many companies have agreed that "data is an asset" concept, but do not know how to count the assets. They see Internet companies based on large data to obtain huge gains, but also thinking how to activate their own large data, but have not found a suitable application, still in combination with their business to find large data profit model.

Caixin Reporter: In your opinion, the future in the big data this industry chain, who will be the main players? What is the trend of its development?

He Baohong: Big Data is extended from the Internet, and all walks of life will flood into big data. Now it seems that the two types of companies in the big data field to occupy the advantage, one kind is the Internet enterprise, for example domestic Baidu, Alibaba, not only grasps the big data technology, itself can capture the user data, has the huge data source, must be able to enlarge the data; A class of companies that specialize in data analysis, they may be small, There is no data, but the tools and techniques that hold large data can be used to analyze companies that do not have large data capabilities, such as in the steel and energy sectors.

The first issue is the openness of government data

Caixin Reporter: As you said, the development of large data is still in the initial stage, then, at this stage of development what problems?

He Baohong: Big Data in the final analysis to have a huge amount of data, now the key problem in the data source.

The first issue is the openness of the data. In fact, the government is the main source of data, if the government's data is not open, the market for large data will be relatively narrow, many innovative applications will not be achieved. As for enterprises, especially the traditional large state-owned enterprises, the data between departments and departments can not be completely transparent and open, it is very difficult to ask them to open up the data.

Therefore, in the primary stage of large data development, we see most of the "private big data", such as the traffic department to master the traffic data, the banking sector to grasp the bank data, the telecommunications sector to master the telecommunications data, and so on, but not shared between each other, forming a "data island."

In addition to open data, standardization of data sources, quality control of data sources are also facing difficulties, the industry is also exploring solutions.

When discussing the problem of data source, there are also a series of data security issues, such as privacy disclosure, trading data and so on. The Ministry of National Information has repeatedly stressed the protection of personal data security, and recently promulgated the "Telecommunications and Internet users personal information protection provisions." However, how to ensure the security of information in a large scale in the age of data, no one knows, because most of the data is not yet open to each other, even have not been linked up.

Caixin Reporter: According to the big data, our government department has promulgated which policy to guide? Is there any relevant policy available for reference abroad?

He Baohong: Big Data is a new thing, there is no specific policy promulgated, but in the government's macro-policy such as "Twelve-Five" planning, has repeatedly mentioned massive data processing problems.

In fact, open government data is undoubtedly the biggest policy support for big data, but it is a gradual process, it will take a long time to achieve real data opening.

In foreign countries, it is also a headache to open data. However, the U.S. government is ahead, U.S. President Barack Obama has made clear that the government information is open, all not secret information must be in machine-readable form to open to the public, such as meteorological data, hospital fees data. Such data openness no longer ends with the publication of a result, that is, the level of information disclosure, but rather the evolution of the data that formed the outcome.

We should be aware that international competition based on big data has come quietly. March 22, 2012, Mr Obama called Big data "the new oil of the future," announcing a 200 million dollar big data investment plan. It can be said that the U.S. government has lifted large data from spontaneous business practices to the height of the national strategy. Under competition, our Government should make new consideration to big data.

Caixin Reporter: What is the value of data openness?

He Baohong: The data is not networked, the value will be greatly reduced. Of course, by analyzing the data of a single enterprise can also achieve some value, but the greater value of the data is that different data sources can be interconnected, like 20 years ago, the computer can be used alone, but once the network, what kind of application, at that time simply unpredictable, The only certainty is that the value behind networking is much better than it used to be.

Large data is the same, we can think that the Internet is now connected to hardware devices, including PCs, mobile phones, tablets, the future of the Internet is connected to a wide range of data to form a data network, the value is not greater?

You can use your imagination, when the traffic department's Road data, the bank's consumer data, the telecom operator's user location data and the Internet manufacturer's goods, these overlap, what kind of possibility will appear.

Caixin Reporter: From the policy level, how to ensure that the data can be opened after the security problems?

He Baohong: In fact, advances in technology have made internet anonymity a mathematically impossible thing. Any form of anonymity and privacy is an algorithmic impossibility as long as there is a reasonable business and security motive. Who you are is no longer important, it is important that you label information such as location, gender, age, interest, direction and occupation.

According to the study, 20 years ago, you can identify 87% people by "sex + postcode + Birth date". And in the big Data age, by analyzing the 4 of users who have been to the point of location, you can identify 95% of users. Large data without original sin, need to reflect on the adjustment is not it, but our own. The big data age needs to adjust our concept of privacy protection. For example, laws and policies should not constantly expand the scope of "personal information" protection, but rather restrict the purpose of large data applications. What privacy protection needs to do is shift the focus of big data regulation from collection to use, not vice versa.