i卡會員

The maturing of 5G technology has increased the demand for edge computing. This, along with the acceleration of artificial intelligence (AI) in different industries, has been pushing up the corporate demand for data, data transmission and computing resources. To keep up with the evolution of emerging technologies such as 5G and AI, data center service providers are faced with the pressure to transform.

The pandemic that emerged in 2020 has been a catalyst for the digital transformation of companies. The adoption of cloud services has become critical to maintaining operations and enhancing competitiveness. Companies’ on-premise data centers have been steadily moving to the cloud. Meanwhile, new applications and services are continually being introduced in the cloud-native environment. The technology focus is on data management, energy and power management, the flexibility of information technology architecture, storage and localized computing capability.

The rise of edge data centers; hybrid IT becomes mainstream

The architecture and equipment of present-day data centers has come a long way since the days of the traditional server rooms. Containerized data centers are currently the mainstream. These are mostly comprised of one or multiple 18-42U containers. Servers, power supplies, air conditioning, wiring and piping is all flexibly configured within the container to facilitate mobility, scalability, replacement and fast deployment.

The rapid development of the Internet-of-Things (IoT) has resulted in the  amount of terminal data increasing sharply. Over recent years, the scale of large data centers (e.g., Google, Amazon and Meta) is becoming ever bigger. Meanwhile, small decentralized edge data centers (EDC) are starting to attract market attention, namely where the deployment and operating costs of large data centers are not aligned with the needs of edge equipment and customers. Going forward, edge data centers will be mission oriented and deployed locally or at specific corporate sites (e.g., micro data centers and modular data centers). This approach saves space and retains the advantages of containerized data centers. Suppliers will mainly focus on customized projects by providing computing capabilities, infrastructure and IT solutions. In addition to cloud services, the applications at the customer’s end will also include additional functions such as cloud + edge, edge + micro and co-location.

Remote monitoring saves manpower and avoids service disruptions

The U.S. consulting and research firm Gartner forecasts the global market for edge computing infrastructure to reach US$17.9 billion by 2025. The investment in optical modules and network applications to upgrade edge data transmission will reach US$1 billion. 

As data centers are becoming more modularized for rapid deployment, the number of edge data centers and micro data centers is set to rise. The corporate demand for co-location will also grow. Given the limited manpower, the operation and maintenance of data centers is a great challenge for both service providers and companies. The application of enhanced AI can help companies to remotely monitor electricity usage, cooling devices, power distribution modules and uninterruptible power supply (UPS) and to record any abnormal events accordingly. This alleviates the workload on employees, as one person may be managing a few hundred micro data centers. Warnings and troubleshooting suggestions will be issued in the case of any faults. Equipment service life can be predicted, so that replacements can be scheduled before problems occur. This avoids service disruptions.  

AI value-added energy management to reduce energy consumption and costs

As the global data volume increases, power consumption has doubled about every four years. In the U.S., the electricity consumption each year is over 90 billion kWh. As 5G, AI and data analytics develop, the requirements for data transmission speed are increasing and the range of functions keeps expanding leading to the demand for more computing power and electricity. Overworked servers in data centers suffer overheating and low computing efficiency. Good cooling equipment is needed to reduce the power usage effectiveness (PUE). In addition, the utilization of AI in server rooms and rack power distribution units (PDU) enhances remote power management and data center environmental monitoring through real-time information from cooling sensors and cooling management systems to maximize the efficiency of the data center.

Many international companies debuted heat dissipation solutions at the Mobile World Congress (MWC) in 2022, as their answer to low carbon initiatives in response to the drive for sustainability around the world. Greater computing power, storage, equipment flexibility, network speed and environmental friendliness were all highlights at the exhibition. For example, Amazon Web Services (AWS) pledged to achieve zero carbon emissions by 2040. The company will launch cloud tools for monitoring carbon footprints for its customers to enhance their understanding of carbon emissions. Huawei has unveiled full-stack data centers emphasizing high computing power, flexible storage, one-stop quick service and energy efficiency. Going forward, data centers will move towards the edge, the cloud, higher density and sustainability.