The cloud computing is an important factor for environmentally sustainable development. If, in the one hand, the increasing demand of users drive the creation of large datacenters, in the other hand, cloud computing’s “multitenancy” trait allows the reduction of physical hardware and, therefore, the saving of energy. Thus, it is imperative to optimize the energy consumption corresponding to the datacenter’s activities.
Three elements are crucial on energy consumption of a cloud platform: computation (processing), storage and network infrastructure. Therefore, the aim is to provide different techniques to reduce energy consumption regarding these three elements using the algorithmic game theory tool.
The scientific community has today the ability to combine different computational resources (possibly geographically spread around the globe) into a powerful distributed system capable of analyzing massive data sets. Cloud Computing expands the previous capability of combining resources from different organizations by making possible the use of commercial computational providers. A resource can be dynamically acquired as “a service” and customers are charged only for the time they use the service. A company which wants to store its data can opt to outsource the storage service required by paying a fee to a storage service provider because the ”rental” expenses may be considerably lower than the total cost of an individually owned data center.
Another important aspect is more general than users’ money saving interests. It concerns the energy consumption reduction in the global context by which we understand all storage/computation servers available in a given network.
We focus our work on the energy saving aspect of these virtualized services on the global scale pursuing the idea of the intensive migration of classical storage/computation systems to virtual ones. We will address this problem by means of its proper mathematical modeling and find an efficient algorithmic solution to it.
This work focuses upon environmental aspects of Cloud Computing which should be taken into account in the context of the sustainable development and the reduction of a carbon footprint. We are interested in distributing customers’ demands in a network which provides virtual computations services. Our main goal is to reduce the consumption of electrical energy. We aim to provide solutions which allow network and virtual service operator to distribute and to schedule users’ task. The subject is to design some coordination mechanisms in order users to have incentive to cooperate for reducing the energy consumption.
We intend to elaborate algorithms of load balancing, which is conditioned by the energy consumption of the entire system as well as users’ requirements.
This work will be divided in five tasks.
1) The problem corresponds to compute efficient solution for several objectives simultaneously. In our context, it is not pertinent to use these techniques (since energy and SLA’s constraints are conflicting objectives). In multiobjective optimization, a tool to capture the trade-off between conflicting objectives is the notion of Pareto set. The first work is to understand how this Pareto set, can be computed.
2) The second work will also focus on cooperation in order to reduce the energy cost. To exploit the recent advances of algorithmic game theory, we are interested to design mechanisms that are able to enforce cooperation among the private clouds. In terms of the game theory we can express two divergent goals realized at each storage/computing server.
Status of funding
Dimanche 15 juin 2014 21:48:40 CEST
Dimanche 15 juin 2014 21:48:40 CEST
Aucun fichier joint à cette fiche
Ecole Doctorale Informatique Paris-Sud
Nicole Bidoit Assistante
Stéphanie Druetta Conseiller aux thèses
ED 427 - Université Paris-Sud
UFR Sciences Orsay
Bat 650 - aile nord - 417
Tel : 01 69 15 63 19
Fax : 01 69 15 63 87
courriel: ed-info à lri.fr