What it Means to Have a True 2N Power Infrastructure
April 30th, 2013
In today’s market place, the term ‘2N power infrastructure’ is tossed around pretty lightly. A term once used to describe a fully redundant power system is now used to describe any number of configurations that may or may not be truly 2N. It’s time someone set the record straight about what it means to have a True 2N power infrastructure.
But before we get into that, let’s start with the basics. Almost all data centers have one or more of the following power configurations available to clients: N, N+1, or 2N. An N configuration simply means that there exists one of everything with minimal redundancy. This typically means one source feed from the electric company, a single power distribution unit (PDU), one battery (UPS), and a backup diesel generator. While this leaves the infrastructure vulnerable to single points of failure, it often works for companies that don’t require an always-on system, and can make due with a single backup generator in areas where the utility power might be unstable.
An N+1 configuration takes this a step further by offering some additional levels of redundancy beyond the generator. A common configuration of an N+1 system would include 2 power receptacles provided to the customer connected to two separate PDU devices, but only a single UPS and backup generator. This type of configuration can often be referred to as PDU diverse. This setup will tolerate basic cabinet and PDU-level downtime, whether they be failures or routine maintenance. So while there exists a level of redundancy built into this setup, it is still lacking the concurrent maintainability of a full 2N system.
On the other hand, a true 2N configuration represents a completely redundant, concurrently maintainable system – meaning that for every component in the power infrastructure, there is a redundant pair. For example, for every generator, UPS, PDU, and power receptacle inside a server rack, there exists a second generator, UPS, PDU, and receptacle that is fed from an entirely separate power system. This guarantees that should one or more devices go offline, it will not impact the successful delivery of power because everything has an online backup. When done right, this is what represents a fully redundant 2N power system.
Where things get tricky is when providers claim to be 2N, but really aren’t. Despite being a commonly accepted definition, I’ve seen many cases where a data center says they’re 2N, but in reality operates at an N+1 redundancy level with a section of their setup that isn’t actually 2N. In order to be truly 2N, there should be two completely independent, fully redundant paths to get power from the source to the destination. As such, in an effort to clear up any confusion, we at redIT have begun referring to our setup as a True 2N power infrastructure, since it clearly represents something that has a backup for everything, which makes it truly 2N.
While it would be nice to believe that most providers actually do that which they say they do, it’s important to visually inspect all of the components in place to be certain that a power configuration will meet your needs. Don’t wait until the power goes out to start asking questions—understand what power configuration meets your needs, and if you require a True 2N setup (hint: I recommend it!), make sure you inspect the infrastructure to ensure that you are getting that which you expect.
-Gabe Faulkner, Facilities Manager