I was sitting around the Crosby family compound in mid-August when the power went out. Power outages are always disruptive—you never realize how many clocks you own until you have to reset them—but they are particularly distressing when it’s summer in Texas. As each minute without air conditioning passes, the level of fear and anxiety rises. Immediately, a sort of survival mentality kicks in and potential alternatives are evaluated. Calls are made to family and friends to develop a potential exit strategy. “See if the power is on at your mother’s.” If that doesn’t work, “Maybe we should go to a hotel?”

It’s a tense time for all involved, and soon things can take on an every man for himself mentality. My neighbor’s kids once got into a fistfight to determine who would be first to fan himself at the refrigerator. But on this sultry summer evening I was ready. After a brief disruption in the ballgame I was watching, the generator I had purchased kicked in. I made the kids reset the clocks, and all was back to normal in the family home.

This experience brings to mind what I call the $500,000 question. Why purchase a dual feed for your data center instead of buying an extra generator?

Certainly the underlying reason for purchasing a second feed is understandable. Redundancy is the foundational concept that data centers are built on. The positive correlation between uptime and customer satisfaction is a concept that all data center operators can embrace. This is part of our data center DNA. But in terms of that second feed, maybe we’ve all become just a little too “reflexive.” It’s become a standard check box item for new data centers, and few consider whether the second feed really adds the value we want.

The prime area of concern has nothing to do with the physical connection when analyzing the benefits of a second power feed. After all, the average time to repair power line cuts is measured in hours, not days. The issue has everything to do with the source that brings power to the grid itself.

The United States power grid is old for the most part, although it is generally reliable. In fact, the grid has outlasted other innovations from that same period, including the leisure suit and polyester pants (the nadir of this country’s attempt to improve fashion through chemistry). Put plainly, the grid is, shall we say, feeling its age.

Since 1991 non-disaster related outages affecting at least 50,000 consumers have increased 124 percent, with over 40 of these incidents last year alone. The northeastern U.S. suffered major blackouts in 1965, 1977, 1996, 2003, and 2011. Even my home state of Texas, which has its own power grid, has warned of having to implement rolling blackouts during the peak of summer. What this means from a practical perspective is that in many parts of the country (yes we’re talking about you New Jersey, San Francisco, and Northern Virginia), the redundancy of your second feed may be based on a less-than-reliable source.

Now I am not trying to bury this country’s power grid, but I am also not here to praise it either. Certainly, as 600 million citizens of India can attest, there are far less reliable sources of power in the world, but the question that must be asked when building a new facility is whether the expense of the second feed is justified by its ability to perform its desired function. Fortunately, the purchase of extra on-site emergency generation offers a viable alternative to alleviate these concerns.

From a financial perspective the choice between a second feed and another generator is clear. Running another line from a sub-station is the type of expense that boggles the mind. I have often thought there is a “too many zeroes” problem when I see the quote from the utility. They charge a one-time set-up fee for each feed and then bill monthly for its use for eternity. This is not to mention the cost to “reserve” the juice as these fees are just to take the delivery and do not include the energy cost. A back-up generator on the other hand is a one-time expense that can be depreciated over whatever period your financial folks desire. It is much easier to make a logical business case for spending $500 to 750K for a 2-megawatt engine vs. the millions spent upfront for a second feed that brings no protection against regional blackouts.

 

The United States power grid is old for the most part, although it is generally reliable. In fact, the grid has outlasted other innovations from that same period, including the leisure suit and polyester pants (the nadir of this country’s attempt to improve fashion through chemistry). Put plainly, the grid is, shall we say, feeling its age.

 

Operationally, the purchase of a second generator simplifies your maintenance operations through its ability to provide you with a concurrently maintainable or fault tolerant environment. Naturally, this makes the scheduling of routine maintenance much easier while providing your facility with the security of a known quantity in the case of a power-related problem. Unlike a second power feed that relies on the effectiveness of the attached power grid, a generator is the source of the power it delivers thereby eliminating the concerns that are associated a dependence on an aging infrastructure.

Thus we have arrived at the $500,000 question. Has the time come to turn against convention and purchase a second generator for your next data center or continue to rely on a power infrastructure that is becoming increasingly less stable? Obviously, you know where my sentiments lie. Just remember that the ramifications of your choice could be a little more inconvenient than having to reset the clocks.