Data center technology is evolving rapidly to keep track of changes in its operating environment. Despite the complexity of the systems installed – or perhaps because of it – many data processing centers use less than 20% of their processing capacity with many servers ticking over unproductively.
These semi-redundant devices still generate heat. This heat then unnecessarily increases cooling load. This short paper discusses three very important things to know in order to manage these energy costs.
- Tracking humidity, temperature and airflow
- Managing physical equipment change
- Enabling reliable delivery of energy and cooling
Tracking Humidity, Temperature and Airflow Using Online Measurement
Having accurate humidity, temperature and airflow data in real time is the first of three very important factors you need to know about data center operations. However, this information only becomes meaningful when we view in terms of the associated computing equipment.
That’s because individual pieces have varying impacts on these dynamics
depending on their ageing, performance characteristics, and actual loading. Hence we need a detailed inventory of what we have by location.
A variety of data center management tools can provide this information. However, those programs using RFID (radio frequency identification tagging) stand out as among the most effective of these.
A Short Overview of RFID Technology
RFID technology uses programmable digital tags to identify individual pieces of equipment, and locate them on a virtual map using radio waves. RFID is therefore similar to barcoding in the sense it stores the parameters on a database.
However, whereas barcoding uses optical scanners that must be in line of sight with the equipment, RFID can read tags almost anywhere in a room. Therefore, RFID is capable of automatically updating itself when equipment moves, or new devices are installed.
RFID technology is a subset of AIDC (automatic identification and capture)
methodologies. All these techniques collect data about identified equipment, and enter this on databases with little or no human intervention.
RFID simplifies this process further by using radio waves to capture information contained in equipment smart-tag integrated circuits.
At headline level the technology comprises:
- An RFID smart label or tag
- An RFID reader with artificial intelligence
- An antenna allowing these to communicate
The reader converts the information in the radio waves into data and transfers this to a host computer. The user can then analyze the information when convenient and make strategic decisions regarding equipment placement.
However, equipment management tools are insufficient on their own. We move on the second of the three foundations to know in order to manage energy costs.
Procedures to Manage Equipment Location in Data Centers
Equipment locations can change in an unplanned fashion during maintenance, emergency repairs, and upgrades. While our RFID may obediently shuffle the information on its database, we do still need to control these physical equipment changes as managers.
Therefore, the second of the three foundations is the need to have structured procedures in place. These should oversee design, deployment, maintenance, and decommissioning of computing and infrastructure equipment in terms of information contained in RFID tags.
We need to become more flexible in terms of data center layout, especially in view of the trend to move critical data to clouds, and essential servers to data colocation centers. Good data design facilitates these changes by providing a flexible environment as follows:
- Power and cooling grids in separate pathways so plug-in resources are
readily available anywhere in the room
- A free air cooling system that reaches throughout the entire space combined with hot and cold aisles
- Multiple drainage gullies, if in a high-flood-risk zone enabling subdivision for other uses in the event some work is contracted out
Flexibility avoids the need to over-engineer a data center for every eventuality. Real power savings are possible when we couple this with modular energy backup.
Enabling Reliable Delivery of Energy and Cooling Using Fluid Dynamics
In physical science, fluid dynamics describes and understands how gases and liquids flow. Engineers use the information to design aerodynamic aircraft, direct mass flows of petroleum through pipes, predict weather patterns, project weapon fission detonations, and even understand how nebulae behave in interstellar space.
On a different level, we can use fluid dynamics to map and improve how cooling air flows through a room, and how electrical energy flows through a grid. This is the third thing we need to know about data center operations.
Having a data map showing the location of processing equipment – and being able to manage physical changes proactively – allows us to fine tune these systems using fluid dynamic software to save operating costs.
A Short Overview of Fluid Dynamic Software
Quality CFD (computational fluid dynamic) software comes loaded with a library of power and airflow features of multiple types of popular IT equipment found in server rooms. It taps into RFID data bases, and information about the physical environment to visually present the energy impacts of future equipment changes.
This information enables the software to predict the impact of such moves on available capacity, airflow, efficiency, reliability, temperature and energy cost without first having to make the physical changes.
Some of the Improvements CFD Analysis Prompts
Computational fluid dynamics analysis facilitates reliable delivery of energy and cooling by following a process:
- Analyze the effectiveness of current rack and aisle cooling
- Identify areas receiving too much cooling air
- Redirect surplus cool air to potential hotspots
- Resolve areas where hot air reduces cooling effectiveness
- Model airflow under raised floors and fine tune flows
- Simulate the impact of computer room air conditioning failure
- Feed this information back into disaster recovery planning
Conclusion: Three Foundations to Know in Action
Tracking humidity, temperature and airflow, managing physical equipment change, and enabling reliable delivery of energy and cooling are like brightening the headlamps on a car so we can see further ahead at night.
The three most important things to now about data center operation we discussed constitute a reliable means of delivering energy and cooling to a data center as effectively as reasonably possible.