Improved Throughput and Response Times by a Light-Weight Load Management Protocol
Today's public cloud computing offerings provide theoretical infinite resource capacity and enable generous elasticity on instances for their customers. Nevertheless, booting new instances means administrative and organizational efforts and supplementary costs. In order to avoid unnecessary additional costs, scaling mandatory resources up and down has to be well-scheduled. Until now, vendors offer only technologies to react to load variations though manual configuration and automatically proactive actions are not enabled. Consequently, possibilities for allocation automatism that considers the differences in dynamics in cloud environments should be further explored and solutions should be developed.