While replicating data in real time does a pretty good job at protecting the data from loss in the event of an array--or larger--failure, replication alone does nothing to ensure that the applications that access that data keep running. If there's a disaster, there had better be a recovery procedure or our apps could be offline a long time. Xiotech, with ISE Continuous Availability, is the latest vendor to come up with a solution that not only keeps the data safe but also keeps it available even through an array failure.
The problem with traditional replication is that the logical volume on the target array, while it has the same data as the LUN on the primary array it's mirroring, doesn't have the same identity as the LUN on the primary array. If your database server has the primary LUN mounted as drive K: holding your ERP database, when the primary array fails, the database server app will crash and you'll have to mount the secondary array's LUN as drive K: before you can restart the database engine.
Hopefully, you've practiced this, written scripts or are using an application recovery tool like VMware Site Recovery Manager. If not, it could take a while to update Fibre Channel zones, LUN masking and the server's mounting of the volume.
Vendor tools such as ISE Continuous Availability, Compellent's Live Volume, HDS' HAM and even EMC's vPlex appliances all allow applications to maintain their connections to data even when an array fails.
Xiotech's ISE storage blades combine the function of RAID controllers and drive shelves in a more conventional disk array. Each ISE has a pair of active-active RAID controllers and 20 or 40 disk drives. The controllers manage the disk drives, including reinitializing drives and disabling individual heads to create a self-healing disk array. Servers are typically connected to both RAID controllers using the server operating system's native multipath software.