2013-03-03

ICT for Sustainability – The Challenge of Making It Real

The First International Conference on ICT for Sustainability (ICT4S 2013) brought together leading researchers from the technical, natural, and social science disciplines to exchange ideas and discuss about an information society that is able to make sustainable use of limited resources. Based on this aim the conference was clustered around three main topics:
  • Sustainability in ICT
  • Sustainability through/by ICT
  • Societal aspects, economic and political dimensions
Within the context of “Sustainability in ICT” we presented our research results regarding the relation between software and its energy consumption. In order to classify software and develop sustainable software, criteria are required. We presented a first approach of criteria to decide if software is green and sustainable or not. The presented criteria are covered by a so called “Quality Model for Green and Sustainable Software” and can be categorized in common, directly, and indirectly related criteria1:
  • The common quality criteria arise out of the well-known quality criteria for software. The model takes aspects into account like Efficiency, Memory Usage, Idleness and Number of Methods.
  • In contrast to that, Energy Efficiency can be ranked against a reference system and goes to directly related criteria. It is the same for Framework Entropy und Functional Types. Additionally, aspects like Hardware Obsolescence, the Carbon Footprint, and Energy Consumption go into this category.
  • The third group contains indirectly related criteria. These apply to software which aims to support sustainable development. Hence, Product Sustainability with its different criteria sums up the effects of software on other products and services.
In addition to that theoretical point of view, Pierson presented existing approaches (mainly software) regarding the “Energy Consumption from a Software Perspective2. These are for example: PowerScope, PowerPack, EnergyChecker, pTop, PowerTop and the so called PowerAPI. Based on that he introduced their proposal of the Energy Consumption Tools Pack2:
  • Energy consumption library (libec)
  • Data Acquisition tool (ecdaq)
  • Data Monitoring tool (ecotop / ganglia plugin)
  • Energy profiler (valgreen)
These tools should support the idea to make application developers more energy aware so that the applications itself become more energy-friendly. Additionally he claims an energy-friendly operation system and middleware.

Koçak presented a case study of measuring efficiency of the DB2 workload of the two scenarios3 following the equivalent target of measuring the energy consumption of software: with and without compression. Together with her team she found that data compression reduces tables’ size by 61% and leads to 97% performance improvement. The slowest configuration of the tested database was the one without compression.

Focusing on “The Impact of Improving Software Functionality on Environmental Sustainability” she pointed out that the challenges of Green IT are immense and advocate to make trade-offs between end-user demands and the requirements for corporate social responsibility initiatives in the context of software modernization3.

Another model within the field of energy aware software was presented by Grosskop. He introduced a model for the “Identification of Application-Level Energy Optimizations4 (developed together with his colleague). With the help of this model it is possible to get an impression of possible energy savings. The idea is to break up the system into sub components to find out the ones which mainly causes the energy consumption of the system. The model contains the following steps:
  • Identifying all hardware components used by the system (node map)
  • Determining the main “units of work” of the application
  • Measuring the energy usage of the components and putting it into the relation to the work done
Analyzing the results it might be possible to estimate the energy footprint of the software system and propose alternative software designs4.

Where the previous speaker mainly considered software application in general or for the end user, the last presentation in this session, given by Van Bokhoven, was about a “Pilot Result Monitoring energy Usage by Software5. During this pilot they monitored four servers at the data center Reasonnet Amsterdam to test the monitoring of processes on servers. They monitored the energy usage of the entire data center and the energy usage per rack and came to some recommendations how to improve the monitoring5.

Overall, the presenters agreed about the importance of energy-aware software and the needed transparency regarding this topic. Furthermore, it is especially important to involve the software developer as well as the end-users.

The proceedings of the conference can be found here.

In my opinion the conference was a well-rounded event. Next to interesting presentations and numerous discussions of the participants the moderator Peter made sure that everyone get involved of the ideas of ICT4S and get in touch with new contacts. This ideas were summarized in a list of round about 40 recommendations "How to Improve the Contribution of ICT to Sustainability". Personally, I hope that these ideas will be improved during future events.

  1. Kern, E., Dick, M., Naumann, S., Guldner, A., Johann, T.: Green Software and Green Software Engineering – Definitions, Measurements, and Quality Aspects. In Hilty, L. M., Aebischer, B., Andersson, G., Lohmann, W. (Eds.): ICT4S ICT for Sustainability. Proceedings of the First International Conference on Information and Communication Technologies for Sustainability, ETH Zurich, February 14-16, 2013. Zürich: ETH Zurich, University of Zurich and Empa, Swiss Federal Laboratories for Materials Science and Technology, pp. 87–94. http://dx.doi.org/10.3929/ethz-a-007337628.
  2. Pierson, J.-M.: Energy Consumption from a Software Perspective, Keynote of the First International Conference on Information and Communication Technologies for Sustainability, ETH Zurich, February 14-16, 2013. http://www.ict4s.org
  3. Kocak, S. A., Miranskyy, A., Alpetekin, G. I., Bener, a. B., Cialini, E.: The Impact of Improving Software Functionality on Environmental Sustainability. In Hilty, L. M., Aebischer, B., Andersson, G., Lohmann, W. (Eds.): ICT4S ICT for Sustainability. Proceedings of the First International Conference on Information and Communication Technologies for Sustainability, ETH Zurich, February 14-16, 2013. Zürich: ETH Zurich, University of Zurich and Empa, Swiss Federal Laboratories for Materials Science and Technology, pp. 95–100. http://dx.doi.org/10.3929/ethz-a-007337628.  
  4. Grosskop, K., Visser, J.: Identification of Application-Level Energy-Optimization. In Hilty, L. M., Aebischer, B., Andersson, G., Lohmann, W. (Eds.): ICT4S ICT for Sustainability. Proceedings of the First International Conference on Information and Communication Technologies for Sustainability, ETH Zurich, February 14-16, 2013. Zürich: ETH Zurich, University of Zurich and Empa, Swiss Federal Laboratories for Materials Science and Technology, pp. 101–107. http://dx.doi.org/10.3929/ethz-a-007337628
  5. Van Bokhoven, F., Bloem, J..: Pilot Result Monitoring Energy Usage by Software. In Hilty, L. M., Aebischer, B., Andersson, G., Lohmann, W. (Eds.): ICT4S ICT for Sustainability. Proceedings of the First International Conference on Information and Communication Technologies for Sustainability, ETH Zurich, February 14-16, 2013. Zürich: ETH Zurich, University of Zurich and Empa, Swiss Federal Laboratories for Materials Science and Technology, pp. 108–115. http://dx.doi.org/10.3929/ethz-a-007337628

1 comment:

  1. It's very important that software companies will also give importance to our environment sustainability while increasing their market value through their software products. Users needs also to be aware the impact of different softwares on their daily lives.

    ReplyDelete