26 March 2015
Title: Accelerating elasticity in clouds
Among the benefits provided by cloud platforms, there are simplicity and agility with which users can update system architecture by allocating or releasing resources on demand. This research aims at modeling the relation between system workload and resource demand. If that relation is known, we are able to predict the amount of resources needed to best accommodate (future) workloads. Taking the Universal Scalability Law as our starting point, we attempt to quickly find the parameters that characterize a given workload, to subsequently find an optimum through elementary curve fitting.