A big data trend on the agenda of the summit

As one of the most watched Hadoop finale of the year, the 2013 Hadoop China Technology Summit is due to open on November 22-23rd at four points by Sheraton Beijing Hotel. The conference assembled nearly thousands of CIOs, CTO, architects, IT managers, consultants, engineers, Hadoop enthusiasts, as well as it vendors and technologists engaged in Hadoop research and promotion, will share the hot topics related to Hadoop.

IDC predicts that in the next few years China will have more and more enterprise users to test water large data platforms and applications, and Hadoop is regarded as the "artifact" of large data analysis, will become the most dazzling "star". Hadoop-related data applications have sprung up in all walks of life. Only take the Internet as an example, at present Jingdong, Taobao, Tencent, Baidu, Amazon, a Amoy, everyone, Youku, Sohu, Sogou, Storm audio and video, Ebay, music, pplive, millet and other IT companies in person fencing, use Hadoop to do a big fight. The so-called times to create a hero, the Big Data era is a heroes era, 10 Committer gathered in the capital, the Hadoop industry hero will be sung 2013 Hadoop China technology summit scene. We've blown up five examples of a fairly representative Hadoop industry to share the highlights of the 2013 Hadoop China Technology Summit. General website: http://www.chinahadoop.com/

Hortonworks to renew the front: Hadoop2.0 strong attack

As the saying goes, "very clever!" 2013 Hadoop China Technology Summit invited the Bald-headed foreigner is the famous Hortonworks company's Asia-Pacific technical director Jeff Markham. Hortonworks, a large data analysis service company, has been working with Yahoo to contribute more than 80% of the source code for the Hadoop backbone project. In addition, Hortonworks is considered a major contributor to Hadoop 2.0, the Apache Hadoop yarn.

One of the major themes of Jeff's speech at the 2013 Hadoop China Technology Summit was Hadoop2--yarn. Speaking of the birth of yarn, Jeff said the jobtracker/tasktracker mechanism of the older version of MapReduce needed to be fixed on a large scale to fix its flaws in scalability, memory consumption, threading model, reliability, and performance. Hortonworks, while embarking on building Hadoop2.0, wants to radically redesign Hadoop's architecture to achieve the purpose of running multiple applications on Hadoop and processing related datasets. This allows multiple types of applications to run on the same cluster efficiently and controllable. This is the real reason why Apache yarn, based on Hadoop 2.0, can be born.

Jingdong's Gold rush in the electric business sector: Leveraging Hadoop for large data areas

Perhaps Jeff missed the "tornado" of China's online shopping, believing he was too late to experience how crazy China's double 11 was. However, if you have the opportunity to participate in the 2013 Hadoop China Technology Summit, you may wish to bask in 11 of those remarkable net buys the gold number, 1.6 million bra above 3 Everest, 9 hour diaper can suck up 6 West Lake. Believe that shrewd he will use this opportunity to find the people of Jingdong to sell yarn.

Internet industry, who has mastered the user data, who has the capital to make money. To Low-cost, authentic licensed and among the electric business giants of the electric business in Beijing East attracted a large number of fans, ten years to accumulate hundreds of billions of valuable user data. Traditional companies such as banks, insurance agencies, telecommunications enterprises, and so on, most of their data are structured, and internet companies such as Baidu, Tencent and other enterprises of data, mostly from the network comments, user logs, the data is unstructured or semi-structured. The data of the electric business enterprise such as Jingdong is in between: from the user to the warehouse sorting, then to the distribution, the data on the whole transaction chain is structured, and the user's website browsing behavior, purchase evaluation and other data are unstructured. What needs to be done is to skillfully integrate structured and unstructured data in order to achieve customer insight, user orientation, risk assessment, and a series of data-related analysis and decision making behavior.

In the electric power merchant Gold Rush, the big data already became the Beijing East Invincible Competition weapon. How to use large data to excavate the data accumulated over the past decade, to provide decision support for enterprises, and to support better and greater development of Jingdong, is the core problem faced by the Beijing East Hadoop team. To this end, 2013 Hadoop China Technology Summit invited 3 senior Jingdong Hadoop Technical experts, from each dimension in-depth analysis of the large data Hadoop applications, including marketing system, advertising push, warehousing, sales forecasts, logistics and distribution. For example, the user does not come to the time of the goods, which does not mean that the user will be there when the goods. Jingdong needs to analyze the amount of user visits and commodity data, integrate a more accurate spot rate, and provide the purchasing department with stock in real time to optimize the user experience.

How do I push the right content to the target customer at the right time? Almost all electric business enterprises will be based on the user's purchase behavior to do precision marketing. Jingdong is no exception, every day will produce hundreds of millions of PV, but its advanced Hadoop data analysis means cover the simple e-mail and SMS momentum. Jingdong relies on large data to model the user and perform the correct portrait analysis and positioning models. To put it simply, Jingdong uses Hadoop to analyze and excavate the user's massive comments and search logs, including gender, age, whether there are many dimensions of the car and so on, and make a large data analysis model to determine whether the user is buying impulse or goal-specific, understand the user's purchase intention, Then, according to different user attributes recommend different products, so as to enhance the user experience, to bring more value to users.

Millet enters cloud service industry: vigorously develop hbase technology

If 58 is a magical website, then Millet is a magical company. By contrast, after a century of old Nokia plus all the patents of total assets is only 7.2 billion, and the establishment of millet valuation produced more than 10 billion dollars. Talk about Millet, its hunger marketing law in the Chinese market is in full swing, even Apple cattle are beginning to hoard millet. Compared to the millet mobile phone, millet large data does not make publicity, but this does not affect its large data in the field of strength. No, HBase's lead man Michael Stack came to see this amazing company. You know, most of the structural data of the Millet cloud service is the use of hbase extension technology storage, Millet submitted 65 hbase patches, of which 37 have been merged into the HBase main code tree. And as the Millet data team, of course, will not miss China's most valuable Hadoop technology feast--2013 Hadoop China Technology Summit platform, to share the scene for everyone how the millet cloud services to use hbase related technology.

Large data storage project director will attend Millet Annual conference

Big data in video games of numbers: Youku potatoes Mining data value with Hadoop

The video seems to be endless, watching a video, there will be one after another related video recommendations, the video industry has become a pioneer in the era of large data. As a well-known large video site, Youku Potatoes has a huge amount of video files. There is a technology that Amazon and Google are using, and Amazon will tell you that "a customer who buys a product also buys a B product", and YouTube, a video playback is over, and the recommended video will appear immediately. Similarly, Youku relies on the "Collaborative filtering recommendation" technology based on Hadoop to give users a video they like to watch.

What is known as the ambition? Of course, Youku is not satisfied with the data mining analysis only used in simple recommendation video, Youku potatoes hope to be able to establish a benchmark in the industry, as its fist platform for strategic products "China Network Video Index" has become a big data era of the tide-goers.

Youku potatoes have massive amounts of data, only operational data, the current daily collection of various types of Web Access logs have reached TB level, after analysis and compression of the historical operation of the data has reached hundreds of TB, will soon soar to a petabyte, 5 years after the amount of data will exceed dozens of PB class. How to better handle and analyze these massive data? How to make nuggets in massive numbers? This will be a worthwhile study for Youku potatoes.

At this 2013 Hadoop China Technology Summit, Hadoop technology experts from Youku potatoes will spot the application of Hadoop in advertising, web sites, wireless, and search. On the Youku potato platform, every time a user clicks on a video, Youku potatoes record the page browsing, comment collection, video playback, and various actions to play. These data after processing the analysis results will be feedback to the different relevant business modules for reference, from products, content operations, user personalized recommendations and advertising business departments will benefit.

In terms of content, Youku potatoes data statistics on user networks: for example, each playback whether the buffer, the average download speed is how much, by virtue of these data for real-time statistics and calculation, access to each region under the user's load performance of each operator, in order to determine the distribution of CDN nodes and distribution strategy, Provide clear and smooth video service for users from different regions and different carriers.

In the recommendation, Youku potatoes through the analysis of a large number of video playback behavior, to induce the correlation between the video of different types and different content, to excavate the homogeneity view habit of users, to make a follow-up recommendation for each user's watch, and to improve the existing service iteratively by the analysis of subsequent data. To provide users with customized push service.

VMWare leads virtualization industry: Hadoop's Big Data extension technology is better

As the banner of virtualization technology, VMware is always leading the development of virtualization and cloud computing. But VMware's ambitions are much more than that. VMware starts to exert the technology of Hadoop virtualization. VMware recently announced the launch of VMware vsphere Big Data Extensions, which will allow the company's popular infrastructure management software to control the Hadoop clusters established by enterprise customers. As a result, thousands of VMware Enterprise customers will be able to use the software they already know to control the Hadoop deployment.

To this end, 2013 Hadoop China Technical Summit Organizing Committee specially invited two VMware Heavyweight technical experts, for you to Sunding, explain VMware's large data program.

Who did you give the investment money to? Who can get their own bucket of gold?

2013 the first half of the IDG, Sequoia People's money to whom? Large data has become the hottest keyword in investment, the Internet to information-oriented, in large data fields, do data analysis, mining and other related technologies favored. It is worth mentioning that 2013 Hadoop China Technology Summit has created a special forum for large data start-ups and investments, such as Sina Weibo fund, American speed of light venture, IDG Capital investment consultant (Beijing) Co., Ltd., Zhong Tong Yintai, star Ring Information (Shanghai) Co., Ltd., Xadoop, Sky cloud data, Cloud creators and other units will share stories about entrepreneurship and investment in big data areas, hoping to help entrepreneurs and investors at the same time. Now the team to buy tickets there are concessions, for more information on the agenda, please visit www.chinahadoop.com/hadoop.it168.com. This year sales as always hot, if you do not want to wait for tickets sold out, please book early.