Ask any data center manager and you’ll be told airflow management is of utter importance. Then why doesn’t it get the attention it deserves? Why is it so important to optimize your airflow? In general, it’s just a small gap in your server cabinet. Can’t be harmful, right?
Let’s start with some basic understanding of airflows within data centers. Nowadays almost all data centers apply hot and cold corridors to optimize the cooling capabilities. In the ideal situation cold air goes straight to the servers’ inlets. The hot air exiting the servers is returned directly to the cooling unit. This setup enables systems to run at the highest possible efficiency, using the least amount of power. The cooling setup has a big influence on the PUE (Power Usage Effectiveness): a lower PUE results in lower total energy consumption of the data center. This indirectly saves the environment and lowers OPEX costs. Could a small gap in your server rack really have that much influence?
As said above, the ideal setup is cold air entering the servers, while hot air exits. Gaps can lead to a higher demand of cold air than actually required by the servers. See it as a large water pipe: it normally needs a specific amount of water. When we make some holes in the pipe, you will need to pump in more water to get the same amount at the end of the pipe.
Another possibility is that hot air returns to the cold corridor, raising the temperature. This could cause cooling problems for the equipment. An example: a typical server requires 24 degrees Celsius of cold air to operate efficiently. The air getting out of the server might be warmed up to 40 degrees. In this example, there’s a gap close to the cold air inlet of the server that causes warm air going to the cold side. The cold air temperature rises from 24 to 30 degrees, and this could lead to problems with the servers.
(two cabinets with similar contents. The left one has good airflow management, the right one doesn’t)
Possible side-effects are:
Server requires more power for its fans to keep the server from overheating
Result: server is using more power than actually required
Server overheats and fails
Result: complete or partial loss of your server, at least downtime
Server’s hardware components are stressed more
Result: Life span of the server decreases
Data centers need to produce more than the required amount of cooling capacity
Result: Additional energy consumption, raising the data center’s PUE
Raise of cold corridor temperature
Agreed temperature SLA could be breached locally
What you can we do to improve airflow management?
Use hot and cold corridors
Always use blinding panels to close unused cabin space
Even though they’re small: close cable cut-outs and other small gaps
Check the alignment of server cabinets: are they exactly next to each other?
Double check your server airflow inlet and exhaust (right airflow direction)
Always (automatically) close the doors between the hot and cold corridor
When removing a rack from a corridor: close the empty position immediately
Using a Cold or Hot corridor doesn’t matter, the separation of both air streams is the important part. Properly applied airflow management will bring forth higher efficiency and will lengthen the life span of your servers.
At datacenter.com we work together with our customers to optimize their airflow management. This enables us to improve the efficiency and performance of the servers, and lower the total amount of power our data center uses.
The picture is an example of one our customers in our Amsterdam data center. They asked our assistance in improving his airflow management. Needless to say, we jumped right in! We connected the customer with a potential supplier. The solution they came up with even saved the customer 60% compared to standard off the shelve airflow management kits! We take pride in bringing different suppliers and our customers together to find solutions for any requirement they have.
Some more information about cooling and airflow in a data center can be found in my other blog.
Hans Vreeburg, Datacenter Commissioning & Engineering Manager at Datacenter.com
Infradax selects Datacenter.com for Colocation
Datacenter.com today announced Infradax B.V. (Infradax) has selected Datacenter.com’s data center infrastructure to supports Infradax vital mission critical services. Infradax is using Datacenter.com colocation services in Amsterdam to deliver Backup-as-a-Service and resiliency services solutions in a reliable, high availability environment.
Obtaining a certification and going thru a certification process costs a lot of time and money. A certification shows you comply to a standard or specific baseline, but why do you need external audits if you implemented that standard or baseline? Let’s walk through the most common reasons.
Datacenter.com announced an expansion of its Amsterdam data center. The expansion will increase Datacenter.com’s equipped space by nearly 50%, will enable the company to meet growing customer demand for high capacity and flexible colocation space and services, particularly from customers in the fast-growing digital media, gaming, enterprise and cloud computing sectors.
Datacenter.com operates large scale flexible data center facilities to meet the market’s growing need for energy-efficient, highly interconnected, neutral facilities, in which organizations can host their critical IT infrastructure. We address the increased demand for cloud computing. The foundation for Datacenter.com is based on making the digital business of our customers successful, by offering custom-made, proven, reliable, secure, innovative and energy-efficient data center solutions, combined with best-in-class customer support.