Who is the "Master Hua" in the data center?

Looking back over 2013 years, IT industry's hottest keyword is undoubtedly "big data". The charm of big data is the huge business value, according to IDC forecasts, China's large data technology and services market will increase from 2011 to 77.6 million U.S. dollars to 2016 616 million U.S. dollars. Large data is like a gold mine, through the appropriate methods and tools, can be mined from the valuable business opportunities, the direction of enterprise decision-making, as well as the way to win competition.

Of course, everything has two sides, big data like a double-edged sword, in bringing value at the same time to bring unprecedented pressure on the data center. Large data has a large number, diversity, real-time characteristics, rapid growth of unstructured data, such as video, pictures, text, voice and so on, a large number of data quickly crammed into the data center, traditional data centers have to face energy consumption, efficiency, cost and other aspects of the problem.

Big Data brings data center challenges

The first problem with big data is energy consumption. According to Gartner Statistics, China's existing data centers 40多万个, energy consumption has accounted for China's total electricity consumption of about 1.5%. In a survey of data centers, 62% of enterprises believe that the data center is facing such problems as heat dissipation, power supply, cost and so on, 23% of the enterprises believe that its data center power supply and cooling capacity is insufficient, restricting the expansion of IT infrastructure, 19% of the enterprises believe that its data center is too large and expensive; Another 17% of enterprises believe that the room temperature is too high, affecting the stable operation of computing equipment, resulting in a series of problems.

It is clear that in the context of rising global energy costs, energy conservation in the data center is the primary issue.

In addition to energy consumption, the low utilization of data center resources is a common problem: a number of third-party agencies, research shows that most of the data center server and network equipment utilization only between 24%~30%, some CPU utilization, hard disk utilization even under 10%, resulting in a lot of waste of server resources. How to maximize the resources of the existing system and optimize the system architecture is the first problem to be considered in the process of transforming the traditional data center to the Green Data center.

There is the problem of operating costs. Large data to let the server continue to increase, the room is constantly expanding, air-conditioning refrigeration is also constantly increasing. As a result, the direct impact of the increase in electricity consumption and management personnel, data center operating costs straight up. Surveys show that more than half of the IT budget in the data center takes place in terms of human costs, with an IT operational budget of 80% to 90%, and duplication of tasks can lead to significant cost-waste and human-error occurrences, bringing intangible pressure to the business.

HP dl380p G8 Solution Puzzle

So how do you address data center challenges in terms of energy consumption, utilization, and operations?

Look at energy consumption first. Data center energy consumption generally by the server and other IT equipment (50%), Refrigeration and Air Supply (40%), UPS and lighting components. The energy consumption of the equipment is the main factor of the overall energy consumption of the data center. Reduce the energy consumption of equipment roughly three directions: through virtualization Technology Integration Server, so as to improve server utilization, reduce space occupancy and energy consumption, has become the main data center of the mainstream consumption reduction technology, the use of high-temperature equipment, thereby reducing the energy consumption of refrigeration equipment; Reduce the energy consumption of the equipment itself and adopt more energy-efficient equipment.

Micro-servers and high-temperature-resistant servers are a good choice for energy-saving and emission reduction in data centers, however, these two special products have some application limitations: Micro-server for those lightweight, parallel application load, such as mobile Web page access, cloud computing, social networks, etc., it is not suitable for high load application load; High temperature server is a new type of server in recent years, it can withstand higher operating temperature, thus reduce refrigeration energy consumption, but need large-scale deployment to have practical significance.

Therefore, the use of low-power general-purpose equipment is more common practice. Take HP dl380p G8 dual rack Server As an example, this server provides a redundant power configuration with a conversion rate of up to 96%, effectively achieving energy saving and emission reduction, and, based on the new Intel Xeon E5-2600 V2 processor, delivers a maximum performance increase of 35% at the same or lower power levels, in other words, With the same load, the server will be more power efficient, and the savings in energy consumption in a large-scale deployment environment would be obvious.

Virtualization is a good way to improve data center utilization, and choosing the right device for virtualization is a key step. Servers that are suitable for virtualization have a high memory capacity and bandwidth, network bandwidth, storage space, such as dl380p G8 with Smartmemory technology, a new memory health plan that supports greater memory bandwidth, fewer DIMMs and less downtime, and memory slots that can reach 768GB of maximum memory With a maximum speed of 1866 MHz, which can increase 16.6% memory performance, while increasing workload performance and energy efficiency, preventing data loss and downtime through enhanced error handling.

In terms of operations, in order to simplify deployment and management, in recent years, one machine product development fast, this kind of integrated software and hardware integration program eliminates the data center deployment and management of many steps, but for many data centers, integrated machine products are too expensive, and often server, storage, network ratio is relatively rigid, And not as versatile as the product with flexible. Select a data center server you should choose more manageable products, such as HP Server integration of HP ILO management Engine, integration of the 4 major innovation, and full hardware integration of all remote management capabilities, virtual media, remote console, virtual KVM and other advanced features; no disc to install the deployment server directly, Intelligent configuration and management, greatly reducing the cost of manual maintenance.

To sum up, if you choose a general-purpose rack server to build your data center, the server's several attributes are not to be overlooked: energy efficiency, virtualization genes, simple and easy to use, with excellent quality HP dl380p G8 is recommended.