James Hamilton has had two interesting recent posts about local renewable generation for data centers that are definitely worth a read for folks interested in “sustainable data centers”: Solar at Scale and Datacenter Renewable Power Done Right.
Sustainable data centers
It’s always interesting to hear perspectives on “sustainable data centers” from industry, because there is a big diversity still in how companies are moving to make their data centers sustainable. Some companies are going with a “local” approach, where renewable generation (in a variety of forms) is integrated on-site, while others are going with a more global approach, where renewables are placed somewhere else on the grid (often nearby, but not always). An example of the former is Apple and an example of the latter is Google.
There are very good arguments for both approaches, and which one wins out for various companies has a lot to do with the workloads of the companies and how directly they are able to monetize their idle cycles. Amazon, where James works, is an example of a company that is very efficient at monetizing their idle cycles (e.g. EC2 spot instances), which has a big impact on their perspective.
But, regardless of the workload, James’ posts make some very important points that are often ignored in the academic community.
First, large-scale solar at data centers is not just placing solar on the roof. The size of the installations needed to go “net-zero” are significant, and this can be a big drawback for the “local renewables” approach, depending on the location.
Second, placing a large-scale renewable installation nearby a data center is often better than having it locally, since it can more easily integrate into the grid control mechanisms, and its use does not place any operational constraints on the data center management.
Using data centers for sustainability
The points in James’ posts really push toward something I’ve been trying to highlight for a while now. The key role of all the workload management and dynamic rightsizing algorithms being developed for handling local renewable energy at data centers may not be the management of local renewable energy. Instead, the key use of such algorithms might turn out to be the integration of data centers into the grid as a large flexible demand resource to ease the adoption of renewable energy into the grid.
People always say that “if we only had storage,” it would be easy to handle solar/wind/etc. But, of course, large-scale storage is too expensive…
Well, data centers are large loads with a significant amount of flexibility that is currently not being exploited! In my opinion, we should see a 20MW data center as a 1-5MW battery that is sitting unused on the grid. (I’ll give the numbers behind this conversion in a future post…but I’m not pulling this conversion out of thin air.)
What we need are new market programs to extract this flexibility, and new algorithms for data centers to manage their participation in such programs… Then, data centers will become a key resource for making the grid more sustainable.
I’ll talk more about the challenges and opportunities for this sort of “data center demand response” in future posts. It’s a big push for our research group right now…