Do dual-corded servers connected to separate uninterruptible power supplies draw power from both?

Dual-corded servers are intended to share the load from UPS systems for overload protection. But do they really work? Expert Robert Macfarlane discusses dual-corded equipment and why data center managers need to be careful when shopping for them.

Yes. Furthermore, something is wrong if the server is not drawing power from both supplies equal amounts. However, there is a caveat when implementing dual-corded equipment in your data center, particularly if you use Power over Ethernet (PoE).

There have been some devices put on the market that claim to be "dual-corded," but in actuality simply have two power cords tied together inside the box that run to a single UPS. This misleading and useless arrangement allowed them to advertise their product as "dual-corded." These products are also potentially dangerous to your data center.

In dual-corded equipment, each power supply is intended to draw power from a different source that is running at less than 50% of its capacity. Both power supplies share the load equally at all times. If power is shut down to one cord, or if one internal power supply fails, the total load is instantly picked up by the other supply, which still has the capacity to handle the total load.

The same principle applies to your power distribution units (PDU) and UPS's. Nothing in the power chain should normally be operating at more than 50% of its rated capacity so it can pick up the total load if its "twin" is compromised. The difference is that you are responsible for keeping UPS and PDU loads in balance, whereas the server manufacturer makes sure the power supplies are designed and rated to accomplish this.

It should be noted that PoE has become something to be cautious of in this regard. The power supplies in big network switches are also load-sharing, but they weren't necessarily designed to handle the load of a fully populated switch with PoE devices on every port. You can keep plugging in PoE devices and the two power supplies will continue to support them and share the load. But they may each be running at more then 50% of capacity. In this case, if power to one side is compromised, the remaining supply will be trying to run at more than 100% capacity.

When running PoE in your data center, it is wise to be certain how switch power is provided, and how much capacity is really available for PoE before power supply redundancy is compromised. This is not something the manufacturers make very clear when selling the PoE option on a switch.

ABOUT THE AUTHOR: Robert McFarlane is a pioneer in the field of building cabling design. He has been asked to speak at countless seminars on building infrastructure for electronic communications, evolving technologies and the requirements of trading floor and data center design. Mr. McFarlane served for twelve years as President of Interport Financial, Inc., a firm specializing in designs for financial trading floors and critical data centers.

Dig Deeper on Data center design and facilities

Cloud Computing
and ESG