Anne-Cécile Orgerie

PhD thesis

Over the past few years, the energy consumption of Information and Communication Technologies (ICT) has become a major issue. Nowadays, ICT accounts for 2% of the global CO2 emissions, an amount similar to that produced by the aviation industry.

Large-scale distributed systems (e.g. Grids, Clouds and high-performance networks) are often heavy electricity consumers because -- for high-availability requirements -- their resources are always powered on even when they are not in use.

Reservation-based systems guarantee quality of service, allow for respect of user constraints and enable fine-grained resource management. For these reasons, we propose an energy-efficient reservation framework to reduce the electric consumption of distributed systems and dedicated networks.

The framework, called ERIDIS, is adapted to three different systems: data centers and grids, cloud environments and dedicated wired networks. By validating each derived infrastructure, we show that significant amounts of energy can be saved using ERIDIS in current and future large-scale distributed systems.

The provided solutions are:

In this context, I developed software:

  • A Grid'5000 log parser (for 2007 traces) written in Perl (1,400 lines) to analyze the platform usage.
  • A Grid'5000 log parser (for traces from 2008) written in Perl (2,100 lines) to analyze the platform usage.
  • A Grid'5000 log replayer (for 2007 traces) written in Perl (3,500 lines) to replay Grid'5000 traces and apply energy-efficient policies.
  • An EARI prototype written in bash (900 lines) for a demo at SuperComputing 2009 (SC09) on the Inria booth. A detailed presentation can be found here.
  • A user interface to access power measurement data of the Grid'5000 wattmeters on Lyon site. This interface comprises a module to obtain the measurements from the wattmeters and store them (800 lines in C and bash), a module to extract past data with a given precision (800 lines in Perl), and a web interface where users can specify their requirements and obtain the wished data. You can find more details here article.
  • BoNeS (Bookable Network Simulator) a network simulator where users can reserve bandwidth, switch off nodes and entire paths, and get the energy consumption of the network. It is written in Python (6,150 lines). See here for more details.
  • The Ecofen simulator, an ns3 module written in C++ to obtain the energy consumption of wired infrastructures and use green functions such as on/off, Adaptive Link Rate and Low Power Idle (work in progress).