For those customers that have existing racks deployed in their datacenters, converged systems vendors provide an equipment list and an assembly diagram that show the end user how to rack, stack and wire the component pieces. Rarely do they tell the IT manager how to power the solution. How many outlets are needed on the PDU? What types of outlets are required? Frequently the networking gear needs C19 outlets, while the compute and storage components of the converged system need only C13 outlets. This requires the datacenter operator to calculate the power requirements for the infrastructure and shop for a PDU that meets the needs of the hardware. An inexperienced IT team may be forced into relying on the plate rating of the hardware components to size the power infeed to the rack rather than using the observed power draw of the component pieces derived from prior measurements taken in the datacenter.
For the adopters of hyperconverged systems, knowing the power needs and selecting a PDU are somewhat simplified because all of the components are the same size, and use the same outlet type. Usually the incremental size of a hyperconverged system is 2U, and the outlet required is a C13. But the datacenter operator needs to plan in advance and make a decision to either provision only enough power for what she needs today, or else overprovision on power and have outlets remaining unused in the rack until the business case warrants implementing additional blocks. Providing enough outlets to allow for a full rack and having enough available power can prove to be a daunting task. Ensuring that rack PDU does not get over subscribed as the rack load is grown takes effort and knowledge. The IT and facilities teams need to work together to ensure that not only is the individual rack able to support additional systems, but also the row level circuit and upstream power systems can take the additional load. This requires visibility to aggregate power draw beyond the rack.