I was asked by David Shroyer at NetApp to moderate a panel on demand response opportunities in data centers at the Connectivity Week conference in Silicon Valley Monday.
We had a great group of panelists, but unfortunately attendees outnumbered us by only a very small margin.
Mary Ann Piette from the Lawrence Berkeley National Laboratory brought the academic perspective, along with the results of an initial opportunity study and the promise of future research into actual demand response participation in data centers.
KC Mares of Megawatt Consulting represented the utility scale data center operator perspective, drawing on his long history of working in this field.
We were also joined by my former colleague Albert Chiu, who works in Pacific Gas and Electric Company's demand response group.
Because we only had a dozen people in the room, the discussion was lively and interactive. I noted a dichotomy of thought: it was postulated that data center operators would participate in demand response programs if they saw a more active, real-time pricing signal from utilities, but at the same time they desire some assurance that their participation would pay off in a predictable fashion so that they could justify implementation costs.
David further muddied the waters by describing how NetApp purchases power on the open market (made possible by a recent easing of direct access rules in California). As expected, NetApp is interested in low cost power, but also price assurance, so they seek long term contracts and have a hedging strategy. In simple terms, they are trying to reduce price sensitivity.
Not to throw cold water on the prospect of data center participation in DR, but a pure market pricing system runs contrary to the desire for pricing predictability. This means that participating in DR may be lucrative one year, and worthless the next, depending on market conditions. That precludes making investments in enabling technology because the business case is unpredictable.
Further, with their very high load factors, data centers are sitting pretty when it comes to real time electricity pricing. Any hit from peak conditions is more than ameliorated by extensive off peak energy use. Other end use customers are more likely to see an overall increase in costs, and will find DR participation more urgent.
We did come to the conclusion that some utility scale data center operators, especially those competing in the cloud service space, are likely to be the early adopters of any demand-side management opportunity, including DR, load shifting, and self-generation.