Does anybody know anything about the OWC Raid-Ready-Enhanced SSDs and the difference in using these for a RAID 0 vs. using the Standard Mercury Pro Extreme SSDs that are not RAID-enhanced?
In a general sense, it is the same distinction that "RAID Enchanced" hard drive have. They are configured for higher than average I/O workload environments compared to the more mainstream. If you put them through a tougher beating then they can take it better. Typically increasing the bandwidth increases the amount and "breadth" of data passing through the storage system.
Primarily what they have is a higher bigger "over provisioning" budget.
It is mentioned in the graphic on this page:
http://eshop.macsales.com/shop/internal_storage/Mercury_Extreme_SSD_Sandforce
Pro -- 7%
Pro RE -- 28%
Simply put, you are buying more Flash storage. ( the controller manages its usage, but it will get used over the long term under demanding conditions. )
The difference means, if the garbage collection algorithms fall behind the curve, the RE version has more room to handle a longer write burst. Similarly, if "burn out" 6-7% of drive in 2 years from writing to the same spot too much then the RE can keep going whereas the Pro drive is spent.
If your RAIDing primarily to read only a constant set of data then there is largely no difference. If write workload is increasing as fast a read workload and want to RAID just as much to keep up with writes.... then 28% over provision makes a difference.
Look at your Read/Write ratios. If write is a very low percentage then may be a toss up. (getting somewhat better wear protection by going RAID 0 since distributing the writes ). if the read/write precentage is 50/50 then RE would be a go. If the write percentage is over 80 and TBs per month or year then may want to pass on both. ( NOTE: I'm sure folks are going to roll in here and claim Flash SSD drives are extremely delicate flowers and try to write them in any non trivial amount they will collapse into a heap... that's not really true if buy a modern, decent one. The over provisioning controls the issue to a large extent for a extremely wide variety of workloads. )
P.S. On younger drives it is doubtful would see a difference unless composed an explicit test to precisely engage the respective over provision capacities. Over time if abuse the drives it should more easily show up. Mainly on differences in write speed.
