Google has called on storage vendors to consider building newer hard drives with a taller design that are primarily intended for use in data centers, abandoning the traditional 3.5” dimension format.

Google’s VP of Infrastructure Eric Brewer, also a professor at UC Berkeley, argues for a revised form factor for hard drives in this research paper. Brewer contends that because hard drive disks are only ever part of a collective in data centers, having larger and more capacious units would minimize potential points of failure and also lead to a more optimized design that abandons the limitations and precepts of today’s standard.

“The current 3.5” HDD geometry was adopted for historic reasons – its size inherited from the PC floppy disk,” Brewer writes. “An alternative form factor should yield a better TCO overall. Changing the form factor is a long term process that requires a broad discussion, but we believe it should be considered. Although we could spec our own form factor (with high volume), the underlying issues extend beyond Google, and developing new solutions together will better serve the whole industry, especially once standardized.”

Brewer claims that altering the circumference of the platter wouldn’t be practical, because while greater width would increase the available storage capacity, it would also bring about a lower Input/Output Per Second (IOPS) because the read head would be required to make a longer journey to the next sector. Alternatively, HDD manufacturers could reduce the width of the platter so that the IPOS would be increased, but this would come at the cost of reduced storage capacity.

As such, Brewer and his team propose that the height of the standard HDD, which is currently one inch for 3.5” disks and 15mm for 2.5” drives, should be increased. That way, more platters per HDD can be stored in the disk, something that would be more economical with regards to packaging and optimal use of the motor actuator and printed circuit boards. In addition, Brewer calls for a newer, smaller HDD with improved seek time to act as a kind of rapid-response buffer between less-accessed storage and the network.

Finally, Brewer also wants the Bit Error Rate on commercial HDDs to be set higher.

“[Since] data of value is never just on one disk, the bit error rate (BER) for a single disk could actually be orders of magnitude higher (i.e. lose more bits) than the current target of 1 in 10 15 , assuming that we can trade off that error rate (at a fixed TCO) for something else, such as capacity or better tail latency,” he writes.

Photo Credit: Jorge Jorquera via Compfight cc