Fujitsu makes storage ‘bullet-proof’

Exponential data growth increases the amount of data hosted by individual storage systems which consequently exposes businesses to a higher impact from planned or unplanned system downtime. The ETERNUS Storage Cluster technology offers application-transparent failover from one storage array to another in the event that one system stops performing. The failover can be triggered automatically in case of system failures, or can be executed manually for maintenance or minor failures. By mirroring data in real time and switching transparently from one system to the other, the data assets are completely protected and business processes are safe from any interruption. There is no need to introduce a storage virtualisation layer.

The adoption of larger disk capacities in a RAID increases the risk of data loss during daily operation. The rebuild time of a disk in a productive system can take many hours or days, and this duration will rise further when new, higher capacity disks become available. Consequently, this increases the probability that a second disk failure can occur within the RAID group before the rebuild is completed, putting data at risk. The new ETERNUS Fast Recovery function uses reserved areas on the disks within a RAID instead of using an exclusive hot spare disk. For this reason write operations can be spread over many disks reducing the rebuild time by a factor of ten. For example, the rebuild time for a 2TB disk can be reduced by a quarter – from 6 hours to 90 minutes.

The huge amount of data in one storage system can cause input/output performance conflicts between various applications. The automated ETERNUS Quality of Service (auto QoS) feature enables storage professionals to assure the needed response time for the most critical business data once the system is under high load. With the new enhancements, administrators can choose to either pick one of three priority levels or set a target response time for dedicated data volumes. If the requested response time cannot be reached through bandwith allocation, the volume is automatically moved to a faster storage tier, for example by using solid-state drives (SSDs).

This is the first implementation on the market where QoS management comprises storage tiering in a fully integrated way. By simply setting target response times, the system undertakes all optimisation steps, complex tuning efforts are reduced to a minimum, and service levels can be tailored to business needs.