A reliable virtual network embedding algorithm based on game theory within cloud's backbone

Oussama Soualah 1, 2 Ilhem Fajjari 3 Nadjib Aitsaadi 2, 3 Abdelhamid Mellouk 1, 2
LISSI - Laboratoire Images, Signaux et Systèmes Intelligents
3 Phare
LIP6 - Laboratoire d'Informatique de Paris 6
Abstract : In this paper, we propose a new survivable virtual network mapping strategy within Cloud's backbone enhancing the Cloud Provider's revenue and dealing with physical failures of routers and links. In order to skirt the exponential complexity of the mapping, we propose a new reliable embedding strategy, denoted by CG-VNE, based on coordination game framework. To do so, we have formulated the problem as two interleaved coordination games. The first game addresses the virtual routers' mapping. In fact, the actions of each virtual router player strongly depend on the mapping of its attached virtual links. Hence, the second game is launched to embed the virtual links. Note that with both games, fictitious players cooperate to reach Nash Equilibrium of which we have proven the existence and it corresponds to a social optimum. CG-VNE aims to maximise the Cloud's provider revenue by maximising the acceptance rate of clients, as well as minimise the blackout rate of virtual networks caused by the outage of substrate routers and/or links. Based on extensive simulations, the results obtained show that CG-VNE has the best performance in terms of i) rejection rate of new clients, ii) Cloud's revenue and iii) rate of clients impacted by physical failures.
Document type :
Conference papers
Liste complète des métadonnées

Contributor : Nadjib Aitsaadi <>
Submitted on : Tuesday, May 22, 2018 - 7:34:56 PM
Last modification on : Thursday, March 21, 2019 - 1:06:52 PM



Oussama Soualah, Ilhem Fajjari, Nadjib Aitsaadi, Abdelhamid Mellouk. A reliable virtual network embedding algorithm based on game theory within cloud's backbone. ICC 2014 - 2014 IEEE International Conference on Communications, Jun 2014, Sydney, Australia. ⟨10.1109/ICC.2014.6883777⟩. ⟨hal-01797869⟩



Record views