Cloud computing environments offer theuser the capability of running their applications in anelastic manner, using only the resources they need,and paying for what they use. However, to takeadvantage of this flexibility, it is advisable to usean auto-scaling technique that adjusts the resourcesto the incoming workload, both reducing the overallcost and complying with the Service Level Objective.In this work we present a comparison of someauto-scaling techniques (both reactive and proactive)proposed in the literature, plus two new approachesbased on rules with dynamic thresholds. Results showthat dynamic thresholds avoid the bad performancederived from a bad threshold selection.
© 2001-2024 Fundación Dialnet · Todos los derechos reservados