How to meet the new challenges of data center facility operation and maintenance in the new era?

Data Center of the New Era

Data centers have only been around for more than 10 years in China, but they have clearly gone through several stages: the first stage (-2005) was the stage of ordinary computer rooms, with UPS power supply, air conditioning and cooling, and IT equipment placed in them was considered a data center; the second stage (2005-10), with the increase in the power of IT equipment in a single cabinet, emphasis was placed on airflow organization, underfloor air supply, and dual-channel UPS power supply; the third stage (2010-15), further optimized airflow organization, closed cold/hot aisles, modular computer rooms, and Tier 3/4 security; the fourth stage (2015-), with the dramatic increase in Internet applications, big data, AI, and cloud services, has led to a rapid expansion and concentration of data centers. Super-large data centers with tens of thousands of cabinets have become mainstream, and the pursuit of energy efficiency and innovative applications have reached a peak. New technologies such as natural cooling, wind walls, underwater data centers, and liquid-cooled servers have been continuously created and applied.

[[222249]]

Current data centers have the following characteristics:

• The scale is extremely large, with more than 5,000 cabinets, and some plans have exceeded 100,000 cabinets; the previous data centers of 10,000 square meters are embarrassed to call themselves big data centers.
• The power consumption is so high that a single 110/220KV substation can no longer meet the power supply capacity. It needs to be supplied from multiple substations. In addition, the power supply voltage is increased, and 10KV power supply is directly supplied to the machine building. There are multiple substations in a data center park.
• The water consumption is large. The use of chillers leads to large evaporation of cooling water, with some units consuming more than 300,000 tons of water per month. The pipe network inside and outside the building is dense.
• There are many new technologies applied at the facility level, including natural cooling, wind wall, liquid cooling, caves, underwater, containers...

New Operation and Maintenance Challenges

In view of the above characteristics of the new era data center, the challenges faced by facility operation and maintenance management are:

• The huge scale has brought about changes in personnel, organization, and efficiency. In the past, manual inspections of data centers within 10,000 square meters took 2-4 hours. Now, with hundreds of thousands of square meters, manual inspections are not enough for a whole day, and smaller responsibility areas must be divided. More operation and maintenance personnel are needed, and the large size of the organization increases the difficulty of management and reduces efficiency. Since operation and maintenance personnel are distributed in different areas, communication between each other is reduced, and they are easily blocked, which makes their mood worse.
• As voltage levels increase, safety risks increase. In the past, maintenance personnel were exposed to low voltage (less than 1000V), but now power supply equipment, generators, and chillers are all powered by high voltage. Maintenance safety requirements have increased, but personnel's safety awareness, work habits, personal protection, and safety education may not all keep up.
• There are many new applications, but insufficient technical capabilities. With the emergence of various new technologies and new applications, there is relatively little training for operation and maintenance personnel, insufficient actual operation and maintenance practice, and insufficient technical accumulation, which will affect the handling effect when problems occur.
• The supply of operation and maintenance talents is insufficient. Faced with the rapid expansion of super-large data centers, the market is unable to provide and meet the demand for operation and maintenance personnel of hundreds of people. However, due to the above reasons, the training and growth cycle of operation and maintenance talents is relatively long, resulting in poaching from each other and competing for limited excellent operation and maintenance talents, which leads to increased operation and maintenance costs.
• The concentration of scale leads to the concentration of risks and the increase of the impact of accidents. A few days ago, an accident at Amazon's data center caused a large-scale service and application interruption around the world, resulting in heavy losses. Therefore, the pressure of operation and maintenance management is ahead of schedule.