Large Data drive data center Change machine Cabinet server landing

With the boom in cloud computing, every company that really knows "data panning" is figuring out a timetable for big data, but is our data center really ready?

In the face of the roaring flood of information, massive data, diverse information services, complex data aggregation and interaction, these future information services, the enormous pressure will fall on the server, whether tower, rack, blade, or machine cabinet server, can provide strong support?

Embracing open source, contributing communities, large data times, there will be more and more users using Hadoop, but is it safe?

Data center changes coming again

Today, more than 85% of international companies are already using social networks, the mobile Internet is growing exponentially in the history of information technology, and information technology workers armed with mobile smart devices use at least 4 electronic devices and numerous third-party applications per day on average. In the future, 2014 will be 90% of enterprises will support enterprise applications operating in personal terminal equipment, 2015 mobile application download will reach 58 billion; 2016 X86 Server Virtualization Application market will grow from 40% to 80%; the market size of the cloud in 2020 will reach 241 billion dollars.

From now to the future, the research data of the major research institutes are explaining a problem, in the cloud computing, the big Data age, the information flood will roar! And behind this problem, the data center will usher in unprecedented challenges, its scale can only passively increase, the resulting space, power, maintenance costs will also remain high. September 2011, Google revealed that its global data center in 2010, the electricity consumption reached 260 million watts, equivalent to a nuclear power plant 1/4, the release of carbon dioxide close to 1.5 million tons.

Looking back, as the elementary particles and core components of the data center, every change in the server product has become the power source for the data center to be reborn again. We have experienced, from the original tower server of High-performance computing, to the rack, then to the Blade server innovation, then, what will be the next generation of high-performance servers?

Server "transplants" and "Change Body"

As the basic unit of Cloud Computing data center, in the diverse server market, the new generation of Ivy Bridge will provide better performance assurance for the user's business system. Compared to the previous generation Cpu,ivy bridge production process and manufacturing process will be significantly increased from 32nm to 22nm. and "transplants" after the next generation of server products will also be the major server manufacturers to a later stage of the main. For example: Shuguang Company's I620-G15 high-performance 2U dual Enterprise Server, as well as the dawning fourth generation 10 core blade server tc4600h and so on.

In addition, because of the different industries in which users are in, cloud computing and information service patterns of large data will be further differentiated, and more and more customized servers will be needed for the diversification of information application patterns. From the future direction of the entire cloud data center, in order to reduce the cost of space, power, maintenance and so on, users will require that the single server must support more users, maximize the energy efficiency ratio, and as easy as possible large-scale deployment and maintenance. The Data Center server architecture will be driven by large data applications, aiming at integrated cooling system, integrated distribution system, integrated delivery deployment, unified integration of cross-platform resources in the direction of rapid forward.

Server because of "integration" needs and "change body", which is the most representative of the famous "Scorpio Project." Under the impetus of Alibaba, Baidu, Tencent, and hardware manufacturers such as Intel, the Scorpio Project-the technical specification of the complete cabinet server solution-took only four months from the first draft to version 1.0. But in the manufacturer aspect, although each big company's research and development department is in the intense preparation, but from the forming commercial product looks, the dawn should walk in front. Due to the early recognition of the advantages and value of the whole cabinet solution to the large-scale server deployment, Dawn in 2010 first launched the Dawning TC5600 machine cabinet Server, and in 2011, "Baidu Arctic Phase I Project" to get the first amount.

In addition, on the eve of the 2013 China Computer Conference, Shuguang company is poised to release the tc5600-h, which conforms to the "Scorpio 1.0" specification, and also serves as a service for cloud computing and supercomputer applications. As a set of independent intellectual property rights of the whole cabinet server infrastructure platform, Tc5600-h compared with the traditional program has a rapid delivery, save space, high energy utilization, integration of many advantages, but also represents the next generation of cloud computing, high-performance computer development direction.

Autonomous and controllable large data platform has landed

In the support system of the data center, whether it is the Internet industry customized cabinets, or large data and cloud computing flourish, high-performance servers to the performance of higher, more dense, more integrated, higher energy efficiency, and the machine cabinet server will conform to this trend of the inevitable outcome.

In addition, in the face of diverse data forms and higher processing requirements, the storage and management of massive data is no longer suitable for the use of traditional data management tools for analysis and processing. As a result, there will be more and more cloud computing center users using Hadoop, and domestic manufacturers need to have autonomic control Hadoop software system. To this end, the dawn for national defense, Government, security and other key areas of the large data platform to provide a large xdata-hadoop software.