Green Data Center: Efficient Intelligence Has Future

The Icelandic and Norwegian governments and agencies have found Steven Sama, vice president of data center services at IBM Global Information Technology Services. He wants to use the Nordic blessed weather resources to build large data centers and provide IT services around the world to accelerate services in Iceland and Norway. The pace of economic restructuring. At present, various large-scale data centers are actively being constructed in various parts of China. According to the latest IDC survey data, by 2010, China has 504,155 data centers, which will maintain double-digit growth in the next few years. By 2015, China will have The data center market will reach US$15.7 billion. The “Twelfth Five-Year Plan” will be a period of great development in China’s data center construction. Will there be a surplus in China’s construction of a data center? After 3 to 5 years, will these data center technologies become out of date? What kind of thinking should we adopt? To build a data center that meets future needs? How should we manage and operate the data center?

Building a data center requires multi-dimensional considerations to build a data center to consider the natural climate, the conditions that are close to the energy, and whether it can continue to attract IT talent.

In the past 20 years, global data centers have maintained a 20% to 50% growth rate, and such growth has occurred in different parts of the world, including the more developed countries of the United States, Canada, and Japan, as well as developing ones. Emerging economies and growth markets. However, from the status of the global data center operations and development recently released by IDC, nearly 80% of data centers have not achieved efficient operation. Of the annual investment in data centers, 70% is for operational maintenance, and only 30% of investments are for innovation.

It should be said that due to the explosive growth of data, all parts of the world have shown a keen interest in the construction of the data center, seeing it as a new bargaining chip for the transitional service economy. However, under new demand and competition, it is not easy to build a data center that can meet current needs, but also take into account future changes, and become competitive and sustainable.

So when Norway and Iceland’s agencies found Steven Sama, he admitted that building a data center needs to be considered in many dimensions, not only considering the natural climate, but also considering the conditions that are close to the energy, and whether it can continue to attract IT talent and IT. Related technical support and other factors.

Shen Litao, general manager of Ordos Yuntai Internet Technology Co., Ltd. revealed that the reason why they established a large-scale data center in Ordos was to see the huge demand for data centers brought about by the explosion of data explosion. On the other hand, it was Yuntai Internet. ZTE Energy, the parent company, has considerable reserves in terms of energy saving, while the average annual temperature of Ordos is 3.5 degrees Celsius. No earthquake occurred and it is also suitable for data centers.

By 2020, the amount of data produced by the world will be 44 times that of today. At this rate of development, the global data center still needs to maintain rapid growth in order to adapt to future needs. But how to build a data center and how to operate it?

According to Yu Wei, vice president of IBM Greater China, and general manager of the telecommunications and media business unit, the starting point and core of the establishment of a data center should be service-based and better shared services rather than scaled infrastructure.

Shen Litao revealed that the main services of their projects in Ordos are concentrated in a few blocks: traditional leasing services; a modular structure that can provide customized services for users; providing targeted research environments for international and domestic cloud computing research; providing data center consultants and consultation service.

At present, many local governments have encircled and launched data center project construction. Competition scale and competition ability should be said to have a certain trend. As Pan Yonghua, the manager of IDC's enterprise systems and software division, said, the development of China's data centers needs to return to rationality blindly, and requires more efforts in business models and sustainable operations.

The data center that meets the current and adapts to the future modular design is currently the mainstream of the market.

How will the data center evolve from now to the future? Will the data center built today be outdated after three to five years?

Steven Sama believes that rapid growth and unpredictability are characteristics of the data center market. Five years ago, IDC conducted a market survey in the United States. Only 11% of customers stated that they would adopt a modular data center. However, by 2009, IDC's survey revealed that more than 80% of customers already plan to adopt a modular data center. Steven Sams also emphasized that modular data center design does not mean that only space and room are modular, but use smaller increments of standardized components to match business requirements with IT requirements and increase data centers when needed. capacity. "I can't guarantee that the data center idea of ​​modular design will be the mainstream of the market in the next decade, but at least for now," he said.

According to Steven Sams, the current data center cooling technology mainly uses air cooling or air cooling, but in five years, water cooling technology may become the mainstream. The allowable normal operating temperature of emerging data centers will also be slightly higher than the current temperature.

Sun Jiangang, general manager of IBM Data Center Service Center in Greater China, believes that if we want to build a cloud data center that can adapt to the present and meet the future, we should meet these basic concepts: full life cycle, forward-looking absorption, multi-dimensional balanced decision-making, and recycling Sustainability and elastic growth.

Smarter operation of data center virtualization is a prerequisite for improving data center capabilities.

It is estimated that the operating cost of a facility is 3 to 5 times the cost of capital in the 10 to 20-year life cycle of a data center, making data centers more efficient and smarter to operate. How to realize the efficient operation of the data center, Sun Jiangang believes that there are four paths to learn from:

One is to optimize IT facilities and maximize storage capacity and availability. He believes that virtualization is a prerequisite for improving data center capabilities. In a joint data center survey conducted by IBM and IDC, 48% of servers in an efficient data center are virtualized. However, at present, only 10% of China's basic data centers use server and storage optimization. The second is to implement a flexible design to support changing business needs. To cope with changing service requirements, data centers need to have adequate availability and redundancy, and more flexible designs are needed for servers, storage and networking, and management planning. The third is to use automated tools to improve service levels and availability. The tens of thousands of data centers are operated and operated worldwide. The reason why they can achieve high-efficiency and low-cost operation and maintenance is due to the extensive use of automation tools. The fourth is to establish a plan to ensure the same business goals. There needs to be an effective planning and mechanism such as technical planning, management and IT planning, and outsourcing planning to deal with changes in the external market environment and within the company.

We noticed that IBM's data analysis acquisition list also began to include data center analysis tools. Recently, IBM has released more than 20 services based on data centers and 28 sets of analysis-based tools, and is adding intelligent analysis elements to its entire line of products. Steven Sams revealed that IBM's corporate command center has to deal with 1 billion alerts per month, and 90% of these alerts are handled effectively through IBM's analysis tools.