CloudBench: an integrated evaluation of VM placement algorithms in clouds

摘要: A complex and important task in the cloud resource management is efficient allocation of virtual machines (VMs), or containers, physical (PMs). The evaluation VM placement techniques real-world clouds can be tedious, time-consuming. This situation has motivated an increasing use simulators that facilitate this type evaluations. However, most reported based on simulations have been evaluated taking into account one specific (e.g., CPU), whereas values often unrealistic are assumed for other resources RAM, awaiting times, application workloads, etc.). generates uncertainty, discouraging their implementations clouds. paper introduces CloudBench, a methodology to deployment strategies private CloudBench considers integration simulator with cloud. Two main tools were developed support methodology, specialized multi-resource (CloudBalanSim), which charge evaluating techniques, distributed manager (Balancer), deploys tests best configurations satisfied user requirements defined simulator. Both generate feedback information, from scenarios obtained results, used as learning asset carry out intelligent faster experiments implemented showed encouraging results new strategy evaluate deploy algorithms