Overview

The energy consumption of the network infrastructure underpinning the Internet is becoming increasingly important. It is estimated to take about 2% of the energy consumption of industrialised nations, or about the same as the aviation industry.

CAIA's GREEN program is investigating ways to reduce this energy consumption.

Activities

The GREEN project is pursuing research that is applicable to the immediate environmental problems facing society, but will continue to be of relevance in the long term. This includes fundamental and rigorous research in resource allocation theory, and grounding in experiments on real systems.

To date, the research in GREEN has two branches:

• Theoretical research into speed scaling. Many devices, such as CPUs and disk drives, consume less power when operating more slowly. This research is characterising the optimal speed to run at as load varies. The aim is to achieve the greatest energy reduction with the smallest reduction in speed.
• Experiments that will qualitatively and quantitatively explore the relationships between IP network traffic patterns and the energy consumption of devices such as routers, wireless access points, content servers and other infrastructure devices.

In the media

An overview of the project presented in the Swinburne Magazine article Internet power may need computers to sleep.

Some background for this project is also provided ABC Radio National's Future Tense programme Solar Roads, urban mining, the Jevons Paradox and energy efficiency aired on 4 November 2010. The section on the GREEN project starts after 20 minutes of the audio download.

Selected Research Output

See also the complete list of publications.

• Better Energy-Delay Tradeoff via Server Resource Pooling (invited)
  Yannis Kamitsos, Lachlan L. H. Andrew, Hongseok Kim, Sangtae Ha and Mung Chiang.
  in Proc. IEEE Int. Conf. Computing, Networking and Communications (ICNC), Maui, Hawaii, 30 Jan 2012. bibtex

• Online Algorithms for Geographical Load Balancing (best paper award)
  Minghong Lin, Zhenhua Liu, Adam Wierman and Lachlan L. H. Andrew.
  in Proc. Int. Green Computing Conf., San Jose, CA, 5-8 Jun 2012. bibtex

• Powering Down for Energy Efficient Peer-to-Peer File Distribution
  Andrew Sucevic, Lachlan L. H. Andrew and Thuy T. T. Nguyen.
  in Proc. ACM GreenMetrics, San Jose, CA, 7 Jun 2011. bibtex

• Balancing Peer and Server Energy Consumption in Large Peer-to-Peer File Distribution Systems
  Lachlan L. H. Andrew, Andrew Sucevic and Thuy T. T. Nguyen.
  in Proc. IEEE GreenCom, 26-29 Sep 2011. bibtex

• Dynamic right-sizing for power-proportional data centers (best paper award)
  Minghong Lin, Adam Wierman, Lachlan L. H. Andrew and Eno Thereska.
  in Proc. IEEE INFOCOM, pp. 1098-1106. Shanghai, China, 10-15 Apr 2011. bibtex

• Greening Geographic Load Balancing
  Zhenhua Liu, Minghong Lin, Adam Wierman, Steven H. Low and Lachlan L. H. Andrew.
  in Proc. ACM SIGMETRICS, pp. 233-244. San Jose, CA, 7-11 Jun 2011. bibtex

• Geographic Load Balancing with Renewables (best student paper)
  Zhenhua Liu, Minghong Lin, Adam Wierman, Steven H. Low and Lachlan L. H. Andrew.
  in Proc. ACM GreenMetrics, San Jose, CA, 7 Jun 2011. bibtex

• Optimality, fairness and robustness in speed scaling designs
  Lachlan L. H. Andrew, Minghong Lin and Adam Wierman.
  in Proc. ACM SIGMETRICS, pp. 37-48. New York, NY, 14-18 Jun 2010. bibtex

• Power-Aware Speed Scaling in Processor Sharing Systems
  Adam Wierman, Lachlan L. H. Andrew and Ao Tang.
  in Proc. IEEE INFOCOM, pp. 2007-2015. Rio de Janeiro, Brazil, 20-25 Apr 2009. bibtex

• Optimal Speed Scaling under Arbitrary Powers
  Lachlan L. H. Andrew, Adam Wierman and Ao Tang.
  in Proc. ACM MAthematical Modeling and Analysis (MAMA), Seattle, 15 Jun 2009. bibtex

What can I do?

As an individual, there are many simple things you can do to reduce your contribution to the Internet's energy consumption.

• Enable the power saving modes of your PC. Most operating systems have a "sleep" mode, from which the PC wakes almost as quickly as from a screen saver. Enabling these modes will allow research like ours to benefit you computer.
• Use a laptop instead of a desktop. Because laptops are optimized to run off batteries, they incorporate the latest power saving technology.
• Shop locally. Even though the Internet allows you to shop for bargains from all over the world, shipping individual items is very inefficient. The majority of energy associated with eCommerce is not due to the networks and servers, but packaging and transport costs. See Energy Use in Sales and Distribution via E-Commerce and Conventional Retail: A Case Study of the Japanese Book Sector by Eric Williams and Takashi Tagami, Journal of Industrial Ecology 6(2):99-144, Apr 2002.

Funding

Lachlan Andrew's work on the GREEN project is funded by an ARC Future Fellowship.

Project Members

CAIA members Lachlan Andrew (Project leader) Tuan Dinh Rozanna Jesudasan Thuy Nguyen Andrew Sucevic Grenville Armitage (CAIA Director) Dragi Klimovski Adam Black

Collaborators Adam Wierman (Caltech) Kevin Tang, Nithin Michael (Cornell) Mung Chiang, Yannis Kamitsos (Princeton) Yoni Nazarathy

Openings for research students

The GREEN project has several PhD projects available.