Home > Research Groups > Algorithms and Complexity > Projects > V:Drive - Intelligent Storage Management

V:Drive - Intelligent Storage Management

Storage Management

Data has become one of the most valuable assets in all businesses, and one of the most volatile. With all businesses, today, the growth of data needed to operate is increasing at the stated 67% per year at current business practices. To worsen the data storage problem, many new regulations (240+ at last count, internationally) require that much data and electronic records and emails be readily available. For these reasons, the requirement for storage continues to grow at a phenomenal pace. Complexity and the proprietary nature of storage systems have meant and continue to mean high management costs. A first step to reduce storage related complexity is the consolidation of storage systems inside Storage Area Networks or SANs. The connection between the servers and storage systems takes place via the high speed interconnects of the SAN. This makes it possible to disband the traditional tight coupling of the servers and storage systems and enables a real any-to-any communication inside the SAN.

Datamanagement

To use the full potential of a SAN, it is necessary to integrate storage management solutions into the storage concept. The SAN management contains different tasks like the administration of switches, remote and local backup of data, and the partitioning of storage systems. An often underestimated task is the distribution of data among the connected storage subsystems. The conventional approach is to tightly couple one or more logical partitions of a disk or RAID system with a file system. This strategy has serious drawbacks concerning the efficient use of the storage systems. Neither the capacity nor the bandwidth of the storage systems can be utilized in an adequate fashion.

The efficient use of a storage system can be significantly enhanced by the integration of a virtualization solution. This software makes a transformation between a logical address space that is presented to the servers and the access to the physical disks. It is possible to use the available physical capacity of the storage systems as one or more big storage pools. Virtual disks are built from these storage pools without worrying about the limitations of the underlying hardware. The aim of this project is to develop efficient strategies for the distribution of data and integrate them into a storage management environment.

Downloads und Links

Contact: André Brinkmann