Intel Senior Vice President Boyd Davis:hadoop Power Support Data Society

The first China cloud Computing conference was opened in June 2013 5-7th at the Beijing National Convention Center. The Conference, with an international perspective, insights into global cloud computing trends, and from the application, explore cloud computing and large data, cloud computing and mobile Internet, cloud security and cloud computing industry applications, such as the focus of the topic. In particular, the Conference has set up a cloud computing services display area, exchange the latest research results of international cloud computing, showcase domestic cloud computing pilot City development achievements, share cloud computing development experience, promote global cloud computing innovation cooperation.

Intel Senior Vice President Boyd Davis

On the second day of the fifth session of the cloud Computing Conference, Intel senior vice President Boyd Davis delivered a speech titled "Hadoop, The force of data society." He describes several major trends in the IT industry today and analyzes the impact of these trends on future procurement behavior in the data center. In his speech, Boyd Davis discussed the use patterns of the Hadoop framework and its reasons for driving the fragmentation of large data.

The following is a live record:

Intel's understanding of large data

First, how does Intel understand large data? We have a lot of experience with CPUs and chips, and we provide processors and chips for PCs and data centers. There is a very interesting trend in the industry, the development of information technology in the 1960 's when the need for automation, led by automation, and then with the development of information technology, the distance between people more and more close, then we began to focus on the problem is no longer to draw people's distance, but focus on the user experience. For example, is this user experience highly customizable or personalized? A lot of big data is generated by the promotion of personalized experience. I believe that the next round of change in the information industry must be driven by big data. So big data is very important for our Intel strategy and our future development, which is why we build large data platforms with many partners.

How do we understand large data? When you talk about big data, you think of terabytes of numbers, or structured, semi-structured data. In fact, when we consider large data in this industry, we mainly consider the tools to deal with large data. For example, relational database, but we found that relational databases have some traditional processing can not adapt to the needs of large data. Big data is not just about data, it's not about processing tools, it contains a lot more business value than normal data. Any country in the world recognizes that data can be changed to produce many business models. The

explores this issue from another perspective, such as Intel's Hadoop, which is very advanced, especially in our market position in China. We used to be a chip company, and now we're starting to focus on Hadoop and some software. We find that the combination of Hadoop and software can make better use of the business opportunities brought about by large data. A few years ago we had a different model of partnering with Chinese partners than it is now, because of innovation. Previous cooperation with China will find a lot of data per year, and the amount of data increased 30 times times, for some operators, if you submit the query needs 30 seconds of time you are not acceptable, 1 seconds can also, Now large data can make these companies closer to their users and provide better service to users. The

Large data also has a significant impact on the intelligent city area. China is developing rapidly, and smart cities are using images to control and monitor video to improve traffic flow and improve traffic, such as Chongqing, which uses data from surveillance to improve public safety or traffic convenience. In Beijing or elsewhere it's a lot different than it was a few years ago, and if there's an intelligent city project going on here, I'm sure the two and three rings will not be as frustrated as they are now. The big data for

is the area of health care. Like people who have cancer,A colleague of mine had a kidney cancer a few years ago with a traditional treatment for cancer, but my colleagues used big data to decipher the secret of the cancer genome. He thought the result looked more like pancreatic cancer, then adjusted the medication, and after a few months he got a good treatment and extended his life. If we can make this a service that is affordable to the public, then it is exciting, such as genetic cracking, and the treatment of various diseases, which is my expectation of big data.

Hadoop cannot solve all the big data problems

Intel's goals: about a few years ago, from Google to Yahoo, Hadoop became an open-source Big Data utility. Hadoop cannot solve all the big data problems, and Hadoop itself is not a so-called solution. But we believe that the Hadoop framework is a very basic framework that enables the application of various structural applications of data on the basis of Hadoop. In traditional industries, a lot of people who can use Hadoop's simpler architecture can be a great help to their work, and we believe Hadoop has great potential to make it easier to deploy and cheaper. And it can be used in a wider range of applications.

So we've improved Hadoop in a variety of areas. For example, real-time analysis is more flexible. Today's Hadoop is a batch processing tool that offers limited value. But now we're going to make it the next platform. It can have a wider range of applications and greater capabilities.

What is the problem with big data now? For example, security, real-time response, the load of the environment, the realization of business value path, and so on. Many companies start by thinking about what data to store and what value can be derived from these data analyses. This is a problem that most companies think about. These advanced companies are indeed considering innovation in value models. But most companies are still thinking about what to do. One reason is because of complexity, because there are many tools that require people to learn the appropriate technology to control them. Some of the analysis we get from the data will really work if we can do it in real time. We have to solve these problems, if the use of highly integrated or vertical base station to do, it may be for many users. But Intel believes that openness can lead to faster and bigger growth in the long term. Closed solutions that are highly integrated or private may quickly solve the problem, but in the long run it is not good for longer term value implementations for large data.

If we have an open, operable foundation or framework, of course, based on Hadoop, it is beyond that to achieve the benefits of the entire industry by implementing services and applications in an open environment.

Hadoop's present and future are different

as we join the Hadoop ecosystem, we are fully committed to the open source. We're constantly contributing to the open source community, and we're leveraging new technology, and Hadoop is, in a sense, a low-end software solution that addresses very complex issues. We can enhance the platform in the hardware layer, this platform to have the security. We can also work with the open source community to drive new projects.

has many opportunities to expand Hadoop to expand it accordingly, and there are many ways to leverage existing technologies to make Hadoop a more versatile application model and system. For example, through the Chi-strong processor can not only improve computing, networking and storage capabilities. We want to make Hadoop's user clusters easier to deploy, especially in terms of storage control. In this way, they are able to intervene after the user knows their data.

We want the Hadoop to be enhanced. Looking at the results, the results are shocking, and some people think of Hadoop as a low-end workload. But without a suitable working environment, it is possible that Hadoop can only be used for low-end work systems, but if it is in a TB environment, the processing speed can be increased by 50%, the exchange rate is increased by 50%, and the hard drive is increased by 40%. The amount of batch processing within four hours of the system can process 1TB of data within 7 minutes.

We want to build an innovative software platform. I said Hadoop itself is not a solution, it is a low-end platform service, most of the value is from the application layer of the upper tier. Most companies have their own architectures, frameworks, and, for example, data mining and analysis of data. Many times it is more complicated to dig up new skills or to ask for a few more, and the cost is higher. With the support of such a Hadoop platform, we can enable more companies to make better use of their data. It is now possible to conduct full-text search and semantic analysis of text-search data, run at the top of the FDS, and improve business value based on text search.

can be seen not only on the application tier but also on the service layer. Most of these big data comes from service based applications, and most companies use Hadoop to provide a range of services. Chinese companies such as Pok and GDS are considering how to expand their relationships with customers in China. Provides a range of services based on Hadoop. For the most part, there is a strong value-oriented approach to doing these things with the expanded, more capable Hadoop platform that we offer from Intel.

Hadoop now andThe future is not the same, for example: (figure) Map reduce data analysis value: Here is a list of different data, we have to go to stereotypes of the analysis processing is very difficult, because the data itself is asymmetric. For example, the person's microblog account has more fans than others, but parallel processing is likely to make certain nodes in the group in an inactive state. Intel has made a technology that uses the organization's data in a graphics-parallel process of Hadoop, which we will launch next year. Not only graphic processing, but also streaming can be used on Hadoop.

Embrace the community and promote open source

Intel is promoting open source in all its aspects, we are now a leading open source software company, we have tens of thousands of software developers, in China there are more than 1000 software developers, most of them with Linux, over the past few years we have a great contribution to Linux. Hadoop is not only a commercial software, but also a source of open source. Open source software is an open tool, but for Intel we develop open source software that is not only able to bring great value and cash flow, but also the ability to advance the entire industry platform. So we want to be able to build the entire ecosystem around the open source platform.

The best solution is based on the best technology, and Intel is a technology company. Our partners can provide solutions for the market, because the storage level, network layer, technology needs of good technology in the end can bring value to the software.

The last thing to say is that you can join us when considering new services or new applications. Because we have one here that can provide a strong foundation for your business growth. Not only can we see large data, but we can see the direction of the future development of large data. If you join us, you can join in a very strong network, in China to help you gather the data needed to solve the problems you face. We have strong hardware or service partners, and we believe that our platform is a very powerful tool and platform to help you move forward. Thank you!

(Responsible editor: The good of the Legacy)