Cloud Orchestration![]() “Infrastructure Repacking” (InfRep) is an innovative solution that contributes to the Openstack project (Nova scheduler). InfRep is designed to be used by all Cloud providers seeking for smart management of their data centers. Using InfRep leads to an optimized physical infrastructure by efficiently reducing energy costs when guaranteeing a high quality of services to cloud consumers.
![]() The main goal of InfRep consists to have an interoperability that provides a dynamic allocation of physical resources and improves at the same time virtual machines migration in a datacenter, reducing therefore the number of working servers. The inefficient and irrational use of physical resources in a datacenter generally creates the appearance of a high number of overload or barely used servers that are due to complex or random demands of virtual machines. As a consequence, the operating and the Capex costs become higher with the continual increasing demands and do not cease to disappear. We propose InfRep as an optimal resource management scheduler, thanks to a new algorithm that makes the urbanization of virtual resources much more efficient and also provides a major improvement for the Openstack project. InfRep creates an efficient VM placement on a minimal and optimal number of physical servers, according to several network, storage and geographic constraints while maintaining a high quality of service. ![]() The interoperability between different service providers has been an important and major problem for a long time in the Cloud Computing domain. InfRep can solve this problem by creating a direct connectivity between several Cloud Computing providers, and even between several data centers that belong to the same Cloud provider. This value eliminates the current “lockup” phenomenon that appears nowadays when a company chooses a Cloud Computing service. It becomes difficult, expensive and sometimes even impossible to change the Cloud provider or to transfer data between different providers or datacenters. InfRep makes possible :
![]() An optimized physical infrastructure consists to manage efficiently arriving workloads often considered as virtual machines, by placing them according to a smart strategy to better use available servers guaranteeing a high quality of servers. After some departures of certain VMs, InfRep is used to propose a repacking plan to optimally migrate virtual resources between servers and data centers. This repacking is based on a fast and scalable algorithm leading to handle with large workloads and infrastructures size. ![]() This value significantly reduces the energy consumption, the number of working servers and eliminates the physical and virtual data storage constraints. The Cloud Computing domain requires huge energy resources: a large amount of electric power for the servers cooling and operation processes. The energy consumption of a server in standby mode approximately represents 70% of its consumption in normal mode. InfRep, with its intelligent VMs placement, uses a smaller number of used servers and temporarily turns off the servers that are not used or not needed at the moment. ![]() The encrypted distributed storage is able to split large volumes of data into chunks, encrypt them separately and then rebuilt and recollect the sensitive data, enabling a secured storage and access. With this value, InfRep provides the Cloud Computing field more secure and ensures a higher confidentiality. The final customer is the only one who can access and use its data. The encrypted data can be stored in different Clouds which are connected using the interoperability feature. The encrypted storage also takes into account all the specific criteria requested by clients, such as: preferences regarding geolocation, data size, SLA requirements, etc. |