Title: energy-aware IaaS-PaaS co-design.
Modern IT infrastructures are big energy consumers. For example, in 2011 Google’s cloud (composed of 900,000 servers) had a total energy consumption around 2.6 TWh. This energy consumption by major cloud was estimated to increase around 15-20% every year. Reducing this energy consumption is an absolute necessity for economical as well as ecological reasons.
Many research efforts have been dedicated to reducing cloud energy consumption, in particular by optimizing the Infrastructure-as-a-Service layer of the Cloud. Infrastructure-as-a-Service (IaaS) is the layer in charge of the virtualization of physical resources, and therefore has direct control over energy-related elements. However, the IaaS layer has no knowledge about the nature of applications which run over these resources, which limits the scope of decisions it can take.
The EcoPaaS project therefore aim at making the IaaS layer (in charge of resources) and the PaaS layer (in charge of applications) collaborate to further reduce the Cloud energy consumption. The idea is to define standard interfaces that allow both layers to exchange relevant information and to coordinate their actions. Exchanging information will for example allow the PaaS layer to estimate the energy consumption of each application it is running. Coordinating actions will in turn allow the system to avoid situations where both layers simultaneously take mutually-damaging actions.
We will therefore investigate difficult scientific questions such as: which information can and should be exchanged between the two layers? What are the necessary coordination mechanisms to control reconfigurations and avoid detrimental side effects between the two layers? And which energy savings can we potentially obtain if we deployed such techniques in large commercial clouds?
We will address these questions in the context of the ConPaaS system running over an energy-efficient private cloud such as Snooze. We will follow a systems-oriented methodology based on rigorous experimental data collection, formulation of hypotheses about the necessary mechanisms, prototype implementations, and experimental (in-)validation of the hypotheses.
- ConPaaS: a Platform for Hosting Elastic Cloud Applications. Guillaume Pierre and Corina Stratan. IEEE Internet Computing 16(5), September-October 2012.
- Energy Management in IaaS Clouds: A Holistic Approach. Eugen Feller, Cyril Rohr, David Margery, and Christine Morin. In Proceedings of the 5th IEEE International Conference on Cloud Computing (CLOUD), June 2012.
- A PhD in large-scale distributed systems, Cloud computing, green computing or high-performance computing.
- Excellent programming skills in Linux environments.
- Good communication and writing skills.
- Good command of English.
- Knowledge of the following technologies is not mandatory but will be considered as a plus:
- Python and shell scripting
- Commercial cloud systems: Amazon EC2 and EBS, Heroku, etc.
- Open-source cloud systems: OpenNebula, OpenStack.
- Revision control systems: GIT, svn.
- Linux distributions: Debian, Ubuntu.
- Note that knowledge of French is not required for this position.
Start date: As soon as possible. The position is offered for a period of 18 months.
Location: Rennes is the capital city of Britanny, in the western part of France. It is easy to reach thanks to the high-speed train line to Paris. Rennes is a lively city and a major center for higher education and research. The job will take place within the INRIA/IRISA research center, which is internationally recognized for its research in the domain of information and communication sciences. The postdoc will take place within the MYRIADS team which is involved in many research projects in the domain of cloud computing and energy efficiency: ConPaaS, Contrail, DALHIS, GreenTouch, HARNESS, Snooze, etc.
Applications: A first selection of candidates will take place on December 13th 2013. To maximize your chances of success, please send your application before this date. The position will however remain open until a suitable candidate is found. To apply, please send the following documents to Guillaume Pierre (firstname.lastname@example.org):
- A motivation letter highlighting your previous research achievements in a domain close to cloud computing.
- Two recommendation letters.