Larger data series Topic Four: What does big data mean to Intel?

If you're familiar with it, you'll know. Intel actually has a lot of free resources to fight for, such as Intel has a software department that can help and guide users in software optimization and development, Intel offers compilers and various development tools, high-performance computing and the Internet are frequent Intel software, and some have joint laboratories with Intel. For example, Intel also has a data center department that provides consulting and services for data center construction and operation, all free of charge. "Intel is designed to help users improve their application level, but given the limited resources, we try to choose representative projects and users." "The head of the Intel Data Center business unit said in an interview. In addition, Intel cooperated with the Education Commission in providing teacher training to China free of charge, and the number of teachers trained in 2007 years reached 1 million. These activities, which are similar in nature, reflect Intel's sense of mission and responsibility as a leader.

Today, large data applications are in full swing. What does big data mean to Intel? Business, Business (life), responsibility and Mission. In my opinion, the latter is greater than the former. If this is the case, it is easy to understand the collaboration between Intel and the Big Data company Cloudera (see: Intel Capital Injection Cloudera fusing Hadoop navigator large data).

No x86, no big data.

It is no exaggeration to say that there is no large data application without x86. A typical feature of large data is low value density, which is why Hadoop was not analyzed for large data because of the scale of inputs and outputs (see:. Business intelligence distinguishes between large data applications). It can be said that without x86, even Hadoop is useless. Large data cores are based on local storage and Low-cost computing capabilities to analyze large amounts of data. Internet enterprises are the first to realize the value of large data applications, Google as an example, the core of the search engine business, is built on the basis of large data analysis, with the help of map/reduse, through parallel computing, Google for the internet massive information, The core of the keyword to build a powerful search function, thus creating a kingdom of Google. Google, with Internet users online behavior analysis as the core, precision advertising marketing has become a portal, electricity and other websites, such as the main business model, which is based on large data applications.

Has the demonstration function of the Internet enterprise, now the big data application spreads widely, compared with the Internet enterprise, the traditional industry scale is bigger, the commercial value is more worth looking forward to. To tap into the commercial value of large data, the best way for Intel is not to rush to the front line and compete with upper-tier partners in the industry, and Intel will be the final winner as long as Intel helps with hardware and software cooperation to make it bigger.

Cooperation Cloudera the RIPE

Intel's collaboration with Cloudera is a ripe one, if not cloudera, it will be another big data maker. Intel's collaboration with Cloudera is primarily a venture-capital act, with 740 million of billions of dollars that is also Intel's largest single investment in the data center sector. In the future, cloudera each revenue will bring a certain percentage of return to Intel, help Cloudera is to help themselves, so Intel and Cloudera cooperation is logical. But the co-operation between Intel and Cloudera is not simply a VC.

It is also learned why Intel chooses Cloudera. Intel's support for Cloudera is not to prop up a firm but to prop up a market. In other words, Intel is not helping Cloudera defeat rivals, which is also Intel's reluctance to see things, Intel's ambitions in industry, which is the relationship between sesame and watermelon. To speak of competitors, Intel's IDH (Intel distribution Hadoop,intel Hadoop distribution) and Cloudera CDH (Cloudera distribution Hadoop,cloudera The Hadoop distribution is the same in positioning, unlike IDH for Intel Hardware partners, offering them a free platform, while CDH is business-oriented, but they are all based on the open source Apache Hadoop platform.

With CDH, Intel is naturally willing to integrate IDH and CDH. It is revealed that there will be a new version of the release in about 3 months. Intel's collaboration with Cloudera will make some of Intel's partners feel worried that the Idh and CDH Fusion will be a commercially available version, and will there be an open source version that does not need to be paid for in the future? In fact, IDH is Intel based Apache A commercially available version of the Hadoop open source platform, Intel stands at a strategic high demand, free of charge externally. Among them, compatibility, stability is the biggest distinction between the two. The Open source platform provides a path to a technology implementation that does not guarantee the compatibility and stability required for business applications. For Internet users, the open source version will be used without hesitation, and they have enough technicians to solve all kinds of technical problems that may arise. But for the traditional industry/enterprise users, there is no such ability, many industries/enterprises are technology outsourcing, the need for third parties to provide technical services. Therefore, for these users, it is more likely to adopt IDH and CDH.

For sure, both Intel and Cloudera will uphold the idea of open source and continue to contribute to the open source community. But as a commercial version, Intel will also help Cloudera expand the market. For existing IDH partners, either rely on their own technical strength, open source version as the core, through the service profit, or around the future version of the integrated business version of the service. It is in Intel's interest, however, that Intel will continue to help partner development anyway.

Why not buy Cloudera

Many people do not understand: Why does Intel not buy Cloudera, it is very simple, for Intel, the big data is the forest, and Cloudera is just one of the trees, no one will give up the whole forest for a tree.

If the acquisition of Cloudera, as Cloudera's founder and chief strategist Mike Olson, said: "If Cloudera is willing to sell, the acquisition means that Intel is directly in line, and today, Cloudera is similar to Intel and Microsoft's cooperation." "As a result, Intel injected Cloudera, the same as the Redhat capital injection."

From the industrial ecological chain, CDH and vsphere some similar, are as x86 hardware platform on top of the management and control platform. In terms of division of labor, Intel focuses on performance optimization, storage, and scheduling, while the advantages of Cloudera are sql/queries, advanced workload scheduling and management capabilities, and therefore, cooperation between the two sides is also a complement of cooperation. In addition, by investing in Cloudera, Intel is also more able to demonstrate its commercial value in the basic research of large data platforms through the commercial market. So it's the right choice for Intel, both in business and industry.

Original source from "Bit net", reprint please keep the original link: http://server.chinabyte.com/184/12948184.shtml