Very high temperatures have baked India this summer, the worst in 122 years. One can roast an egg on a rock. Pune, known for its moderate climate, feels like an oven. Its high temperatures now but there can be cyclones and flash floods in just three months from now. Frequency and intensity of natural disasters has surpassed all limits. All of us know carbon emissions due to energy consumption is the main culprit. We have been warned by none other than the United Nations that we better limit the temperature rise to less than 1.5 degree Celsius (2.7-degree Fahrenheit) by 2030 else our mother Earth will not be habitable for living beings.
Unfortunately, computers have played their part in the deterioration of the climate. A laptop used for 8 hours a day generates 30 kg to 90 kg CO2 equivalent. A desktop computer generates 3 to 5 times CO2 equivalent as compared to a laptop. The Internet is not far behind. Massive data in Cloud meaning data centers of web-scalers like Google, Amazon, Microsoft and others means more carbon emissions due to energy consumption. Even mobile phones and charging of the same left ON can deteriorate the climate since half the world’s population has mobile phones today.
Computers used by people, corporations, and governments also power the Internet and thus create a massive carbon footprint. Who drives computers? Software! So along with hardware designers we the software people have primary responsibility for designing systems that are energy efficient and not energy guzzlers. Here are some simple guidelines for us:
- In the Systems Life Cycle methodology, introduce review steps from the standpoint of reducing energy consumption in computing and network traffic. Sustainability is added as a criterion in addition to reliability, scalability, modularity and zero defect (if one ever achieves that).
- Semiconductor chipsets that drive technologies such as artificial intelligence consume power. Unless serious machine learning is needed, its best to keep applications simple. Not using AI may seem counter intuitive, but the rationale that the application does not justify it and it is environment friendly will find resonance with end users.
- Network traffic goes up for centralized or cloud hosted applications. While Cloud is an accepted norm today, data and computing at the edge has gained currency. A distributed architecture that reduces data traffic will be preferable from an energy efficiency standpoint.
- All wireless devices have WiFi and Bluetooth radios. Mobile phones and some tablet computers also have cellular radios. These radios consume power. State of the art has reached a point where radios go into 'sleep' mode if there is no activity and wake up when needed. Waking up algorithms need to be smarter. At times, WiFi turns on when a familiar or say known WiFi is sensed. Its got to be smarter than that and when these devices are on an application, the environment can figure out if the radios need to be in 'sleep' mode or awaken. To reduce power consumption not just to extend the battery life but from sustainability view point.
Let’s count the number of trees we need to plant for our technology usage. Simply put, each person needs to plant at least ~14 trees over the lifetime to be carbon neutral personally (CO2 emission at individual level) and ~2 to 5 trees every year for the use of laptop and a mobile phone.
