Macroscopic modelization of the cloud elasticity

Jean-Matthieu Etancelin (1) , Sylvain Faure (2) , Tony Fevrier (2) , Chuyen Huynh (3)

1 EDP - Equations aux Dérivées Partielles
2 LM-Orsay - Laboratoire de Mathématiques d'Orsay
3 Cisco Systems France

In order to achieve its promise of providing information technologies (IT) on demand, cloud computing needs to rely on a mathematical model capable of directing IT on and off according to a demand pattern to provide a true elasticity. This article provides a first method to reach this goal using a "fluid type" partial differential equations model. On the one hand it examines the question of service time optimization for the simultaneous satisfaction of the cloud consumer and provider. On the other hand it tries to model a way to deliver resources according to the real time capacity of the cloud that depends on parameters such as burst requests and application timeouts. All these questions are illustrated via an implicit finite volume scheme.

Mots clés: Cloud computing; Modelisation; Finite Volumes

Domaines: Equations aux dérivées partielles [math.AP]

Jean-Matthieu Etancelin, Sylvain Faure, Tony Fevrier, Chuyen Huynh. Macroscopic modelization of the cloud elasticity. ESAIM: Proceedings, 2013, 43 (CEMRACS 2012), pp.136-146. ⟨10.1051/proc/201343009⟩. ⟨hal-00936569⟩