After describing the major IT components in a data center and their power consumption, we presented how servers are cooled and the challenge of air-cooling dense servers with the trend of increasing power of processors and accelerators leading to higher heat capacity per rack. We presented various cooling solutions for the data center leveraging air cooling and water cooling and how waste heat can be reused to cool the IT devices in the data center. We showed, through the recent evolution of processors and accelerators, that higher performance is still delivered every generation, but at the expense of an increased power consumption, an increasing number of cores and the introduction of more specialized processors.

We analyzed the power, temperature and performance of a server when running various workloads. We showed that the two major application characteristics to influence its power consumption are CPI (cycle per instructions) and GBS (total memory bandwidth), while the processor frequency is the major system characteristic. We analyzed the impact of cooling on power, temperature and frequency, and showed water cooling allows processors to run at lower temperature due to the high heat capacity of water, enabling servers to either consume lower power or run at higher frequency, leading to potential higher performance.

As accurate measurements of the of the IT devices power and performance is critical for the data center, we presented the low-level sensors and APIs to measure the server power and performance, and compared their accuracy and granularity. We showed different techniques exist to model power and performance for predicting the impact of a frequency change on the server energy or of a cooling change on the data center PUE. We also presented different software to manage and optimize the system energy or the data center operations.

After presenting the example of an air-cooled and a water-cooled data center and how they can achieve low PUE and ERE, we analyzed and compared PUE and TCO of various cooling designs. We analyzed the TCO of these various cooling designs under various circumstances (existing or new data center, electricity price, free cooling ratio, server power load). Introducing ERE in the TCO calculation, we showed that waste heat generated by hot water-cooled servers and reused by adsorption chillers to produce cold water to cool the air-cooled IT devices in the data center can reduce the data center energy by 40–50% and has a tremendous impact on TCO, with a payback ratio between 15% and 25% over an air-cooled design making this solution highly attractive even for existing data centers.

We presented how renewable energy through PV and wind turbines can be effective to produce the energy still required by the data center, leading to reduced CO₂ emission and toward a net-zero energy data center. As renewable energy production is highly variable, we presented how hydrogen electrolysis and PEMFC can store the excess energy and retrieve it when needed. Although these technologies are still being prototyped, we have seen the technologies are available and can provide the data center energy through a local grid without the losses of high voltage AC lines. In this self-sufficient data center running on DC power, AC to DC power conversion is also becoming useless improving again the energy efficiency and simplifying the data center infrastructure.

Looking back to a legacy data center as presented in Figure C.1 where electricity losses are numerous and where 100% of the heat generated by the data center is going to the sky, we described available and affordable technology improvements to reduce the data centers energy consumption and carbon footprint.

Figure C.2 shows such a data center running smart software to reduce the energy need through the integration and optimization of the building and the IT management infrastructures with the job scheduler and system resource manager. Hot water is used to cool the dense servers and cold water with RDHX for the remaining IT devices. The hot water is reused to generate through adsorption chillers the cold water for the RDHX; renewable electricity is produced and stored locally, and used when appropriate to cover the remaining data center electricity needs, leading to a net-zero energy data center.