Power Redundancy Planning
When it comes to maximizing network uptime and creating a mission critical environment, there are six aspects of power redundancy that need to be taken into consideration.
The first aspect deals with the incoming utility power sources and main power panels. If one power source fails, the second source is still delivering power to the facility. While every datacenter is designed differently, QuadraNet’s Los Angeles datacenter currently has two power grids supplied by two separate four Megawatt LADWP Feeds.
The second power redundancy component is backup generators and automatic transfer switches. These are ready to activate if the incoming power feeds fail. Power is automatically transferred to the backup generators and maintains continuous power to the datacenter. For example, QuadraNet’s Dallas datacenter has six backup diesel generators that are tested bi-weekly.
The third level of power redundancy involves uninterruptable power supplies and maintenance bypass panels. QuadraNet datacenters are configured in either N+1 and 2N configurations. “N” represents the amount of UPSs need to keep the facility powered, and the +1 or 2 designates that there is either one extra UPS or two times amount of UPSs necessary. In these configurations, the datacenter always have more UPS capacity than the amount needed.
As the power makes it journey from the utility supply to the servers, the fourth level of power redundancy planning takes place at the power distribution unit or PDU level. The PDUs serve as a bridge from the building’s primary power sources to the equipment racks. In some cases, the PDUs are sub-panels from the UPS.
Up until this point, the first four steps are designed by the datacenter and should be evaluated during the pre-sales process. From this level forward the rack administrator has some decisions to make on how they want to design the power flow into their rack for equipment usage and redundancy. It’s at this level that most common power failures occur.
Once we are on the rack, the datacenter PDU’s connect to the fifth layer rack level PDU’s that deliver power to the individual equipment in the rack. In a fully redundant design, racks are configured with two independent PDUs each fed by separate facility level PDUs. Careful planning is required at this level to decide how much power is drawn from each PDU so that in the event of a failure the second rack-level PDU will not be overloaded and shut down the rack.
This leads to the sixth and final component of a redundant power plan, dual power supplies on an equipment level. In a mission critical rack, each piece of equipment should contain dual power supplies with each connected to a separate rack PDU. This goes beyond the power supplied to the building and rack failing by also protecting mission critical equipment from a hardware power supply failure. In the event of failure, the second power supply takes over and supplies 100% of the necessary power to the equipment until the faulty supply is replaced.
Meticulous planning and experience are required to engineer a rack that is capable of withstanding power issues at every level discussed. QuadraNet has a team of network engineers that are experts is creating mission critical rack plans. Organizations interested in creating a fully redundant power plan should contact the sales team and schedule a consultation with one of QuadraNet’s datacenter engineers. You can contact sales via email at firstname.lastname@example.org or call 1-888-5-QUADRA.