netapp has moved from zcs to bcs disks for a reliability reason netapp won't change back to zcs for performance reason
bcs doesn't provide less storage space because they are formatted to include additional sector for block cheksum in adition to the existing storage. as a proof of that, consider 18G and 72 G disks : they can be inserted in a bcs or zcs volume because they are slightly larger than zcs drive (36 G drives are an exception)
bcs drives doesn't really impact the Filer performance because - i think - the block checksum is controlled at the disk level (I hope so, as it should be) this block should be harwarelly controlled (as the flow control in a modem can be hard) so the perf impact is related to the disk throughtput and this is the disk manufacturer concern
i think for such file server as netapp , reliability is a crucial point and should override the too-small perf increase of using zcs over bcs drives (if any)
steve@hq.newdream.net wrote:
I'm looking for any type of data that supports whether BCS drives perform better/worse than ZCS drives. In particular, I'd like to see benchmark read/write performance on F760 filers (using ONTAP 6.3.1R1) with FC9 volumes of purely 36GB ZCS or 36GB BCS drives and also volumes of purely 72GB ZCS or 72GB BCS. I looked through the NOW pages but couldn't find anything along these lines. It seems netapp has moved in the direction that all their newer/larger disks tend to be BCS(520) only or am I mistaken? If thats the case then where is the data to show why BCS is better than ZCS. Aside from the fact that you loose a slight amount of disk space on BCS drives what are the pluses? Is there less CPU overhead one way or another?
Thanks,
Steve