Green Networking: Redes Sustentáveis – Prioridade Real ou Discurso Institucional?

Summary

Camila Silva’s seminar argues that green networking is a practical necessity, not merely corporate sustainability rhetoric. The presentation says the digital sector uses about 2% of global electricity and produces CO₂ emissions comparable to those of commercial aviation. It also warns that sustainability claims can amount to “greenwashing” when companies’ practices do not match their public commitments.

Main Technical Problem

Networks and data centers are commonly sized for peak demand but often operate well below capacity. The seminar cites typical server CPU utilization of 6–12% and network or data-center utilization of 5–25%, leaving equipment consuming power while underused. It also identifies mobile antennas as a major source of telecom-network energy consumption.

Technologies and Approaches Presented

• Virtual-machine consolidation: Combines workloads on fewer physical servers to raise utilization; the presentation suggests utilization could reach about 70%.
• Dynamic voltage and frequency scaling: Lowers processor power use during periods of low demand.
• Free cooling: Uses external air or water to reduce reliance on energy-intensive cooling systems.
• Power-down and sleep modes: Put idle equipment into lower-power states.
• Carbon-aware computing: Schedules flexible workloads for times when electricity is forecast to be less carbon-intensive. The seminar cites potential emissions reductions of 1–8%.
• Energy-efficient networking: Keeps unused links idle and uses a proxy network interface to handle simple requests without waking an entire system.
• Lightweight cryptography for IoT: Uses less computationally demanding encryption to reduce CPU use on constrained devices.

Measuring Efficiency

The seminar discusses Power Usage Effectiveness (PUE), calculated as total data-center energy divided by energy used by IT equipment. A value closer to 1 indicates less overhead. The presentation gives a historical average of 1.83 and says cloud computing and virtualization can bring it closer to 1.1–1.2.

It also describes Carbon Usage Effectiveness (CUE) as a measure of direct and indirect CO₂ emissions per unit of equipment energy use.

Trade-Offs and Example Application

Potential drawbacks include delays when waking sleeping systems, possible data loss, bottlenecks if workloads are over-consolidated, and the variability of solar and wind power—which can require batteries or grid backup. The seminar also notes that machine-learning tools used to forecast and manage networks have their own energy and cooling-water costs.

As an example, a smart city could use lightweight IoT protocols for sensors, place unused network routes into sleep states overnight, and schedule non-urgent cloud processing around periods of cleaner energy while keeping emergency services available in real time.

Overall Conclusion

The presentation says green-networking techniques can reduce total network energy use by 27–42%, with performance remaining above 98%, according to figures cited in the seminar. It concludes that sustainability should be considered when systems are designed, especially as AI data centers increase demand for electricity and cooling water.

Reviews, Guides, or Tutorials

None; this is a seminar presentation rather than a product review or how-to guide.

Main Speaker and Source

Camila Silva presented the seminar, which is based on five cited articles described as mostly IEEE publications.