A power outage, even if just for a split-second, would result in hours of downtime, possible loss of data and significant manpower to restore all of the hosted servers to their pre-existing state. This is why our provider has invested over $650,000 invested in power equipment.

Automatic Procedure in the Event of a Loss of Utility Power
Our provider's automatic transfer switch (ATS) monitors the flow of electricity through thier own transformer from the local utility provider (Kansas City Power & Light). In the event that the ATS detects a loss of electricity, the entire load is placed on the batteries. Because the electricity is constantly passing through the batteries, a loss of utility power does not require any sort of "switch" and a loss of utility power automatically places the power load on the batteries.

Immediately after detecting a loss of utility power, the ATS communicates with the diesel generators, and requests that both start. If one fails to start, the ATS knows not to use it; otherwise it will select one of the two. Once the generators are fully running and ready to accept the power load, the ATS then switches from the utility power (which is not working) to the generators. In the meantime, the batteries are continuing to provide all power needed. The process for the ATS to recognize the loss of utility power and to switch the load to the generators takes between 6 to 9 seconds.

Once the utility power comes back on-line for 30 contiguous minutes, the ATS then transferrs the load from the generators to the utility feed, and again places the power load temporarly on the batteries. The batteries recharge once the load is placed back on the utility feed.

Other Considerations
The loss of power from the utility is not the only way to have a potential loss of power. A failure and/or cut in one of the busbars that supplies the power from the basement power facilities to the 10th and 11th floor Internet Data Centers (IDCs) is also possible, though quite unlikely. Therefore, our providers invested in dual busbars which run on opposite sides of the building, and continue to supply power in the event one is cut or otherwise fails. Hardware failure is also possible, and is the reason for the N+1 redundancy in the 400kVa Liebert Series 600 systems.

As you can see, our provider's power failure procedure was carefully planned, financially justified, and is a testament to the commitment of quality of service. Needless to say, our providers have never suffered an IDC wide power failure. We believe their power failure procedure is far beyond any investment of most other SME web hosting firms.