Cost-effectiveness in the cloud - resource allocation in private IaaS clouds

Hartung István <>
BME Irányítástachnika és Infromatika Tanszék

Goldschmidt Balázs dr <>
BME Irányítástachnika és Infromatika Tanszék

Cloud systems may be attractive for smaller or newly formed companies due to the significant amount of cost that can be saved on hardware purchase and configuration and maintenance costs. The same is true for companies that provide services which are periodically loaded but unused in most time of the year. Unwanted hardware allocation may also be caused by certain software configurations, where the individual elements require to run on different operating systems, but never use the full capacity of the given system.
Most private cloud services lack the ability of automatic resource allocation and distribution, which is especially true for open source IaaS systems, where only the basic scheduling and allocation algorithms are implemented, which do not provide real energy saving. Our research focuses on virtual machine migration and the management of hardware allocation in various open source IaaS clouds. Our goal is the development of these systems to enable unused resources to be switched in energy-saving mode. I will present the results of this research, covering many different cloud providers current state.