Data rewrite the future! Interview with Yahoo Research and development Center Han Yiping

"Yahoo has more than 90% of the value of http://www.aliyun.com/zixun/aggregation/13873.html" > Data Driven "-Yahoo Beijing Research and Development Center Senior manager Han Yiping

When it comes to data, Han Yiping is quite excited, he told reporters, inside Yahoo, the concept of data has been deeply rooted. Our engineers are not software engineers, but data engineers; almost all of our products are data-driven; more than 90% of Yahoo's value is driven by data.

▲ senior manager of Yahoo Beijing Research and Development center Han Yiping

More than 90% of the value is driven by data. Such numbers are enough to scare some companies that still ignore the data. To learn more about Yahoo's use of data to create value, Yahoo's "Data platform" progress, and how companies should use Hadoop to realize data value mining, IT168 reporter interviewed Yahoo Beijing senior research and development manager, The fifth session of the Hadoop China Cloud Computing Conference, the joint Chairman of the Program committee Han Yiping.

One, the current situation of Hadoop "in the ascendant" the prospect of deeper and wider

The outlook for Hadoop is different from the industry, and in Han Yiping's view, it is eight words-the ascendant, the broader.

IT168: What do you think of the development of Hadoop and its future application prospects?

Han Yiping: The current situation of Hadoop can be described in a very appropriate word, that is "in the ascendant."

It's been almost four years and nearly five years since we first made a Hadoop salon in China to the fifth session of Hadoop in Chinese this year. 08 Salon, only some enthusiasts or interested people to participate in the first time in 09, the main play is basically Yahoo, Facebook, some of the big U.S. companies. Of course, there are Baidu, China Mobile to do this work. So by the end of last year, a lot of companies have appeared, the name basically does not come up, basically Chinese internet companies, the larger companies are already used, even including other industries of small companies. From this year's registration, more companies will join in this year.

We also see that, in the early days, many companies just came to know what Hadoop was and what it was about. Then slowly more and more companies are coming over, is to understand exactly how I should use, in the end how, I can start to use Hadoop, I can participate. More than ever, through the use of the future, more and more company people come up with some ideas, problems and experiences in use, and then ask how to improve Hadoop.

So, why do I say it's in the ascendant? Despite these years, Hadoop has developed a lot, but it can be said that the future market will be bigger, more companies will participate in, already in use companies need to have more in-depth use.

The application perspective of Hadoop can be divided into several directions: the first direction is from the landscape, we will have more applications, more and more applications, such as Yahoo has been from the first search using Hadoop, developed to the current Yahoo most products are using Hadoop.

Vertically speaking, on the one hand, the future in addition to Internet enterprises, there will be more industries to enter. In the United States there are already many banks that already use Hadoop. In China I also hear that there are a lot of data-intensive companies in the banking, power and communications industries, and they are beginning to understand the use of Hadoop, which I think is a direction; On the other hand, the application of Hadoop becomes more and more in-depth From the beginning, we do some simple experiments, off-line data processing, and slowly become large-scale data processing, online product analysis and so on.

IT168: Now that a lot of business companies are in the circle of Hadoop and some companies are launching commercial versions, does that mean that Hadoop is in a new phase? Have you entered the business circle from the academic session? Does this have a certain impact on the development of the open source community?

Han Yiping: First of all, Hadoop is never just a scholarly thing. The origins of Hadoop began as a commercial application, and the earliest Doug began to do Hadoop and soon joined Yahoo, where the initial development of Hadoop revolved around a very important business application--Yahoo's web search--and then slowly the actual application of other companies, It has never been a research project and has been a very commercial and practical project.

Some companies have been making commercially available versions from 2009 or even earlier, and more are doing things such as assistive tools, packaging, solutions, training, etc. These things have had a great effect on the popularity of Hadoop, as the initial use of Hadoop requires a lot of time to learn about Hadoop, and even a lot of knowledge about the system, distributed computing, and so on to do its development and deployment.

The advent of these commercial versions makes the application of Hadoop much simpler, making it possible for companies without strong technical backgrounds to be able to apply them. More importantly, it is particularly important that when they are having problems, especially some relatively simple questions, someone will be given some direct support. These commercial versions can be said to have Hadoop become a commodity from a technology.

This has a positive effect on the development of the community, because there are more users, it can mean that this thing has more opportunities for development, but also to get more feedback.

On the other hand, there are special people to do or support Hadoop, so that it itself to promote some of the problems, can be better solved, such as the promotion process no one answered questions, not enough documentation.

(Responsible editor: The good of the Legacy)