Computer networks comprising hundreds, even thousands, of servers require ever greater performance in the storage environment. Ultimately, CPUs want feeding. What could be more obvious than to transfer the scale-out in the compute to the storage?