You are provisioning a new 42-U enclosure for a cluster of twelve dual-power-supply servers with an expected maximum draw of 400 W each. Two independent 30 A, 208 V branch circuits (A and B) terminate at the rack. The solution must provide redundant power paths so that loss of either circuit does not shut down the servers, allow remote per-outlet power cycling from the NOC, and conserve all 42 U of rack space for IT equipment. Which rack-PDU deployment BEST meets these requirements?
Feed a single 1-U metered horizontal PDU from a 3-U static transfer switch that accepts circuits A and B.
Install one metered zero-U vertical PDU on circuit A and one basic horizontal PDU on the same circuit, leaving circuit B unused.
Install two switched zero-U vertical PDUs-one fed by circuit A and one fed by circuit B-mounted on opposite rear rails.
Install two basic 1-U horizontal PDUs fed from circuit A and daisy-chain them behind the servers.
Using one switched zero-U (vertical) PDU on circuit A and a second switched zero-U PDU on circuit B is the only design that satisfies every requirement. Zero-U PDUs mount on the rack's side rails, so they do not consume any rack units reserved for IT gear. A switched model supplies per-outlet metering and remote on/off control, meeting the NOC's manageability requirement. By feeding each server's two power supplies from different PDUs/circuits, the rack keeps running if either circuit fails, providing true path-level redundancy.
The other proposals fall short: basic horizontal PDUs lack remote switching and occupy valuable rack space; using two PDUs on the same circuit eliminates redundancy; and placing a static-transfer switch upstream still leaves only one horizontal PDU and loses 4 U of space, while also failing to offer per-outlet control.