Data centers have become a crucial, growing part of enterprises’ production and management systems and a valuable tool to promote competiveness and operating efficiency. But, as the demand for more information plus applications based on Big Data, cloud computing, and the Internet of Things (IOT) continue to grow, traditional data centers are becoming large and unwieldy. The need to up-scale or out-scale data centers makes construction and management more difficult and complex, and enterprises face unprecedented challenges to their conventional methods and processes.
Meanwhile, the modular data center, comprised of micro modules, is becoming popular. The modules’ flexibility and predictability are being used in large telecom and Internet companies, as well as large- and medium-scale enterprises.
Most recently, manufacturers have introduced the “intelligent micro module,” which fuses the Internet of Things and digital and information technology. With it, independent, non-thinking components such as power, cooling, and IT infrastructures become smart, intelligent, and manageable. They also can communicate with each other. Because all systems are visible and manageable, they can be analyzed and optimized from a more comprehensive viewpoint.
Let’s look at the evolution of data center technology and processes and see how intelligent micro modules will result in simpler, more efficient and reliable data centers.
Challenges of Traditional Data Centers
Traditional infrastructures are failing to meet ever-changing business needs. A major reason is that the service life of individual data center components varies so widely. For example, the working life of an Uninterruptible Power System (UPS) is about 10 years; for air conditioners, 10 to 15 years; and for distribution facilities, sometimes more than 20 years. Because it is so difficult for enterprises to predict future business and data center needs, they often over-build initially. This strategy inflates initial capital investment, brings down load rate, raises energy consumption, and prolongs return on investment.
Mounting energy use shrinks profits and influences global climate change. Globally, annual data center power consumption has reached 300 billion kW/h, which accounts for about 1 percent to 2 percent of the world’s total, and is equivalent to 100 million tons of coal. Huge electricity bills erode corporate profits and increase the economic burden of the enterprise. Currently, the average data center PUE of western countries is about 1.8, while in China, the average PUE stays between 2.2 to 3.0.
Labor-intensive, inefficient operation and management also lead to waste and a poor ROI. With the increase both in scale and complexity and the rise of energy and labor costs, traditional methods of data center management and operation are unsustainable and need to be changed. Specifically, traditional data centers lack: a) automatic, intelligent methods to analyze the status of data; b) forward-looking preventive measures; c) efficient operations; and (d) improved rates of asset and resource utilization.
Finally, traditional data center management depends on the redundant multiple facilities rather than trying to improve operational reliability. As a result, actual loads may be much lower than design capacity, creating low-efficiency operations and overwhelming increases in operating expense. For this, blame the lack of real-time monitoring of environment and device status and health.
Popular Micro Modular Data Centers
The modular data center, comprised of micro modules, has become more and more popular. The modules’ flexibility and predictability are highly valued by large telecom and Internet companies, as well as large- and medium-scale enterprises.
Google and Facebook have set up large, modular prototype data centers for outdoor use. These prototypes are enclosed in containers and use cold/hot aisle containment enable modular data centers to provide fast deployment, energy savings, and low PUE. However, in regions short of available land, especially in eastern China, these models cannot easily be copied. The indoor modular data center, which is more appropriate in a multiple-story building, is gradually becoming popular in China. Micro modules have been widely accepted by China Mobile, China Telecom, and Internet giants like Tencent. Their flexibility and predictability also are popular with large- and medium-scale enterprises.
Leading manufacturers have been moving forward to the next step, which integrates the concepts of “intelligent” and “smart” into the new micro modular data center solution. What’s more, some countries have planned to bring micro modules and intelligent micro modules into national or industry standards to improve the efficiency of energy and management.
‘Intelligent’ Data Centers
The construction and operation of data centers need to transition from rigid to flexible, rough to precise, and labor-dependent to tool- and process-dependent in order to create intelligent data centers. These data centers feature:
Fast deployment, easy duplication and online expansion that help reduce initial investment and ensure maximum return on investment.
Energy conservation of non-IT facilities and reduced PUE through highly efficient power supply and cooling. Green IT architecture and optimized operation of every subsystem and device also reduce enterprises’ cost of energy and PUE.
Digitization, which makes systems visible and manageable. Component, assembly, and system levels become detectable and measureable, which facilitates data center management and controls the information of all devices.
Centralized, automatic, and intelligent management and maintenance, which can provide one big, efficient platform for high throughput as well as the ability to analyze Big Data. The traditional data center — with numerous networks and distributed monitoring and management — makes it impossible to share information between operations and O&M teams, resulting in exponentially increased OPEX.
Networked design, with its multiple-layer, precise energy management, results in highly efficient operations: energy savings, asset life cycle management, and optimum capacity management.
Advantages of Intelligent Micro Modules
Simple design, rapid deployment
Intelligent micro modules inherit all the features of micro modules — independent functional modules that meet different business requirements with flexible configurations. Creating complete data centers takes just weeks, instead of the months required by traditional data centers.
Proactive O&M for higher reliability
Reliability is the central issue in data centers. For financial and Internet industries, especially, any service interruption may lead to tremendous losses. The Intelligent Micro Module solution proposes an innovative concept of proactive O&M to monitor, in real time, key, vulnerable components such as batteries, capacitors, air-conditioning fans and valves, and then generate a health assessment report. These measures will greatly improve the reliability of the data center.
For example, by continually monitoring batteries’ resistance, temperature, voltage, current, and other key parameters, the intelligent controller can respond immediately to alarms and take quarantine measures if the data changes out of normal operating ranges. The controller also combines factory parameters, historical discharge data, maintenance records, and other relevant data for deep analysis. In this case, it can predict possible failures from subtle changes, and recommend precautions.
Another example: a key UPS component is the bus capacitor, which can actively determine its health through real-time monitoring to avoid equipment downtime induced by sudden failure. The Intelligent Micro Module combines sensor and Big Data analysis techniques, and changes traditional fault maintenance into failure prediction and proactive maintenance.
Intelligent operations management system
The Intelligent Micro Module can provide simultaneous, efficient operations in order to provide the best green data center, the best asset management, and optimum capacity matching.
A variety of management system reports pinpoint additional energy consumption and, based on large data analysis, provide energy-optimization solutions. This helps to create a green data center. For example, the traditional air-conditioning system always determines how to work according to its out/return air temperature and humidity; there is no unified management to make the air conditioners coordinate with each other. However, with an Intelligent Micro Module, all air conditioning is managed by the centralized intelligent controller, which issues instructions based on a temperature field comprised of temperature and humidity sensors.
A few manufacturers go further. A real-time temperature and humidity field map can be generated by the chip temperature in server and storage devices and inside sensors that collect temperature and humidity readings. As a result, potential cold/overheated areas can be accurately located. Accurate cooling can avoid situations in which humidifiers and dehumidifiers run simultaneously at the same location. Accurate cooling can, therefore, reduce PUE by 0.07 to 0.15. This is a great boon in locations where free cooling cannot be applied.
The Intelligent Micro Module’s built-in, modular UPS can maintain efficiency at about 96 percent, even at a low load rate. At the same time, the intelligent module maintains communications with IT devices through the smart controller, which can manually or automatically output orders to migrate a business system to improve the rate of equipment utilization. The controller also can order idle equipment to remain dormant, reducing power consumption by more than 80 percent.
This concerns the life cycle management of data center assets, including automatic identification and data capture, asset change management, routine maintenance, and the end of an asset’s life cycle. Standardized ITIL service management enables closed loop management of asset information, and ensures the timeliness of data refresh and management. At the same time, based on the precise monitoring of assets’ life cycles, it helps users create maintenance plans that include proactive alerts. It can dynamically adjust the maintenance plan; optimize schemes according to the current situation; and build the best asset management capabilities with Big Data technology.
The Intelligent Micro Module promotes optimized deployment of data center capacity.
Precise management of building- and cabinet-level capacity is based on real-time monitoring of power, cooling, and space. By keeping records of historical operating data analysis, the module can accurately estimate actual demand; reduce the margin between actual and potential demand; and achieve precision matching of power, cooling, and load. Based on a precise algorithm and “what if” simulation test methods for power, cooling, space, and ports, it analyzes and tests the potential impact caused by changes. As a result, it obtains the best cabinet usage from installed power and cooling systems and maximizes the use of installed capacity.
With continued upsizing and centralized development of data centers, traditional data centers must change from an inflexible structure, inefficient management and operations to a digitized, networked, intelligent data center in order to maximize investment returns and operate a networked, intelligent structure. The intelligent data center is the best choice to achieve these goals, and will occupy a vital position in the future.