Selecting and installing your first UPS

As an IT professional, you see UPS (Uninterruptible Power Supply) just about everywhere. Every server room will likely have one of these somewhere. Datacenters are filled with banks of them. Yet there is not much that is written about these boring, yet essential, pieces of silent (literally!) hardware.

Drawing on my own experiences, I’ve come up with three simple steps for the newbie making their first UPS purchase.
1. Size up your battery capacity

The first step in any UPS installation would be to first size up your requirements and identify the servers that you intend to connect to the UPS. While trying to match the correct UPS model to your server isn’t that complicated, it can be overwhelming if you are not a hardware type of person.I will detail how to compute approximate power requirements in a later article, but my advice to the novice is to pass on this part to your vendor for their advice. Some manufacturers like APC will gladly field such enquiries and give you exact recommendations if your servers are fairly mainstream brands, and you are able to furnish them with the model number of your servers.

For simplicity’s sake, I group UPS into three installation “sizes” — small, medium, and large.

• Small: Low-capacity UPS that is normally used to power a couple of desktops or a single server. These can be easily purchased off-the-shelf.
• Medium: Any UPS that is designed to serve more than a single server, yet still plug-and-play. These have to be purchased from IT equipment vendors.
• Large: Any UPS that cannot be directly powered off of an unmodified wall socket is classified as “large.” Also, unless you have prior experience with electrical matters, you should get the engineer from your UPS manufacturer to come in, or else engage the services of a qualified electrician. These UPSs tend to be available only from the manufacturers themselves, or from select vendors.

Most server rooms will probably consist of a number of medium-size UPSs. They might run independently, or are cascaded with battery packs to prolong runtime. I am generalizing here, but two or three of these is usually sufficient to support a rack of equipment.
2. Determine the best location

Unless you have a very good reason not to, you should always install your UPS at the bottom-most section of your server racks. The primary reason is that UPS are heavy beasts, and it makes sense to place them as low as you can.

Some UPSs come with adjustable rails, allowing you to easily slide the UPS in and out of your server rack. Should you use them? Well, feel free to go ahead if your vendor offers free installation. Otherwise, my advice is — don’t bother, assuming that you can replace the UPS battery even when it is mounted directly into the server rack.

I have not experienced any particular benefits from having the rails. It could mean a lot of time wasted trying to get them adjusted correctly. This is especially true if you are doing it yourself, have not done it before, or are just not confident. In these scenarios, just install it in place without the adjustable rails.

3. Prepare for installation

A very important point to bear in mind is that you should never attempt to install a UPS by yourself, unless you are handling a simple, small UPS. The weight of most UPSs is substantial.

Even if you can carry your eight-year-old on your shoulder just fine, you must remember that the weight of a UPS is condensed into a relatively small package with an extremely poor weight distribution — the dense battery is normally located in one end, and it has hard metal edges on most sides to boot.

In addition, keep in mind that you are dealing with electrical equipment, and there is always the chance of an accident. While you might like the peace and quiet of installing it over a weekend or late at night, to be safe, make sure there is someone else around.

Have you been in charge of acquiring and installing UPS before? What are your experiences with it?