– Mark Harris, vice president of marketing at Racktivity (www.racktivity.com), says:
PDUs are required for every rack to deliver multiple outlet power ports. Every Rack in a data center must have PDU devices. The number of outlets is based on the number of servers and switches and similar devices installed in each rack. Typically this is 16-24 devices, so the most common PDU sizing is 24 outlet ports. The topic of interest recently regarding these power distribution devices is the level of intelligence required for each given customer. Some customers treat these as mechanical devices only, others expect these power devices to establish monitoring and control points at the device level. With the industry in turmoil due to the increasing cost for energy, the majority of data center operators are looking for a much more intelligent power distribution device, with deeper insight into exactly how and where energy is being used. The higher the intelligence of a rack level PDU, the more granular a data center operator can be when analyzing their overall costs, and establishing their unique optimization initiatives. All operators are looking for ways to trim costs. Energy is a huge portion of their total costs, perhaps 40% or more today. The opportunity to trim energy costs requires a deep understanding about current usage and trends, at a very granular level.
Since energy in the data center consumes at least 40% of the overall data center costs, everything that provides an opportunity to actively manage these costs should be prioritized at the TOP of the list. Power costs will continue to rise, and anyone not actively looking for visibility and control of power is missing their opportunity and negatively impacting their future.
There are so many choices and price ranges. For years these were treated as commodity insignificant devices, so the culture to look for advanced rack-based PDUs has not existed for very long. There are tremendous choices for highly intelligent PDUs, and these come with higher costs than their much less capable siblings, but the higher cost devices are the ONLY way to obtain a granular working knowledge of energy consumption in the data center. Highly intelligent PDUs are available today that combine quality and intelligence at relatively low prices when considering the total cost of a rack and it’s included active devices. The bottom line is that full racks may have a total cost for all components exceeding $100,000 or more, and the difference between a poorly made PDU device and a highly quality, highly intelligent PDU device will be less than $1000! Why risk power issues and limit power visibility to save what amounts to 1% or less?
Include higher performance PDUs in all future planning cycles. The budgets must be allocated to allow intelligent high quality devices. Look for high quality vendors of PDUs who have technical differentiation and understanding about data center power. IT managers should raise their personal awareness to the energy challenge presenting itself over the past several years to become just as versant in power as they have traditionally been in servers, storage and networks. Ask for help from PDU manufacturers. There is a wealth of knowledge available about power and power monitoring. IT managers can increase their personal understanding, and then implement new approaches to power distribution and analysis within the data center when coupled with the proper tools.
Data Center operators are looking for ways to manage energy in support of their higher-level business management goals; Where abandoned servers are located so they can be decommissioned, predicting critical resource devices that are going to fail in the future which are going to affect availability and uptime, etc. An effective and affordable means to cut energy costs is to move to online cloud storage.These type of predicative analytics are ONLY available using Racktivity’s Genius PDU devices.
Consider bringing all of their power distribution devices into a single manageable domain. Look for ways to combine the energy performance, metrics and KPIs from ALL portions of their enterprise and then actively compare and contrast the operational excellence across physical and logical domains. Looks for improvements that have been deployed in some places but not others, etc.