Original Paper: Rubio J.C.R., & Ramos, M. C. (2018). Data Center Heat Distribution Modeling Using Onboard Sensors.

Frustrated with unresponsive social media pages and delayed emails? Losing online games due to lag? Sometimes, it’s not our devices or internet service providers (ISPs) that are at fault. Data centers, which store, process and manage the information that we consume, have an important role in giving clients speedy services.

A data center is a facility that keeps a huge amount of storage servers and networks often running 24 hours in 7 days. Services such as data storage, backup and recovery, data management and networking are part of a data center’s functions. This means that the contents of our emails, file attachments, videos, avatars for online characters, are being stored and accessed in their respective data centers. As users and processes increase, one of the many issues needed to be addressed is the thermal management of the data center.

DOST-ASTI's RDC thermal map for the top and mid layers

Data centers often experience overheating due to uneven cooling of the setup. To address this problem, temperature inside the facility should be maintained using an appropriate thermal management strategy. This will help avoid thermal imbalances between the hot and cold aisle system of a data center. Maintenance of servers usually involves a hot and cold aisle system with a recommended operating temperature of 18°C to 27°C based on the ASHRAE (American Society of Heating, Refrigerating and Air-Conditioning Engineers) thermal guidelines. However, such a system cannot produce a perfectly desired cooling environment due to the occurrence of heat recirculation which happens when some of the hot air from the hot aisle recirculates back to the server inlets cooled by the cold aisle.

Researchers from DOST-Advanced Science and Technology Institute (DOST-ASTI) and Electrical and Electronics Engineering Institute (EEEI) presented a study wherein a customized or Custom heat distribution model is to test its efficiency in thermal management strategy. The model was first implemented in DOST-ASTI’s Portable Mobile Data Center (PMDC). The Custom heat distribution model generated a more accurate thermal map compared to other heat distribution models. Utilizing the custom heat model provides ease in distinguishing server areas where immense heat comes from providing efficient data center cooling maintenance at a lower input cost and electric consumption. In addition, these advantages will improve the data center’s thermal management strategies by providing vantage relocation of servers and efficient implementation of hot-aisle containment. The system also delivers competitive results by allowing real-time thermal monitoring via onboard sensors.

Amid the rapid development in networking technologies (e.g. Fiber connection, 5G network), research such as DOST-ASTI and EEEI’s Custom heat distribution model should be cultivated. This will provide a flexible and long-term solution to thermal management issues in data centers brought about the increase in users and data processes. After all, advances like this are critical as it is imminent that the majority of Filipinos across the country will have access to high-speed internet connection and will subsequently require fast responses from data centers to provide them the requested data.