In a previous blog I highlighted the idea that not all data center locations are created equal and that there may be a ‘best execution venue’ for every workload type. That trait was one of the motivations behind a recent European Commission funded research project that I was lucky enough to be involved with - RenewIT
The main aim of the three-year RenewIT project was to develop software tools and research to make it easier for data center operators, and their customers, to make informed decisions about where to build new capacity based on access to renewable energy and overall efficiency.
The project ended up developing a range of software and research including its own renewable-optimised data center infrastructure management (DCIM) tools. However the main output of the project was the web-based RenewIT tool.
The free to access web-based tool enables its users – including HPC-specialists – to essentially specify a data center equipped with a variety of energy efficient technologies and on-site renewables. Users are then able to compare how efficiently that site design would perform in more than 50 city-locations across Europe from Reykjavik to Athens.
The RenewIT tool is unique in that there are other sites for comparing access to renewables and also tools – such as Schneider Electric’s Trade-off tools – for modeling different data center efficiency technologies but not one tool that does both jobs.
RenewIT is especially useful as it allows up to five different scenarios to be compared. For example, a facility using free cooling in London can be compared with the exact same configuration in Reykjavik. But a third site, using free cooling and on-site solar panels, in London could also be modeled and compared with the other two models.
The tool provides comparative metrics including PUE, operating and capital costs, as well as the renewable energy ratio (the proportion of renewable energy used by the site).
For example, the image below shows the results of a comparison between a 1MW site in London (Scenario 2), running HPC workloads, with a rack density of 10kW and equipped with free-cooling, versus the same specification facility in Reykjavik (Scenario 1). The calculations are based on an operational lifespan of 15 years (this can be adjusted in the tool).
The results of the comparison show that while the capital costs of building in either location may be similar (there may be more variation in reality), there are big savings in primary energy (non-renewable energy) use from locating a facility in Iceland compared to London. This, as well as other efficiency benefits, results in operational savings of more than 600,000 Euros per year. (The PUE values may be higher than expected in some circumstances but the reasons behind this are explained on the tool homepage).
Another unique feature of the tool is that as well as being able to model the benefits of grid and on-site renewables, it can also be used to compare a variety of different energy efficient power and cooling technologies including free-cooling, heat-reuse, fuel-cells, combined heat and power systems, and energy storage (see image below). The tool also enables the benefits of application-level energy efficiency measures such as workload consolidation to be measured.
The tool has been used by a number of companies including Netherlands-based financial services company ING, which was able to model the benefits of using biogas fuel cells at one of its data centers. Other notable organisations that supported the development of the tool include CBRE, Schneider Electric and the US National Renewable Energy Lab.
The judges at last year's Data Center Dynamics Global Awards which took place on the 7 December recognised the potential benefits of the tool and nominated it as a finalist for the Best Data Center Initiative of the Year. Why not start 2018 off with sustainability in mind and check it out: http://www.renewit-project.eu