Brahim, you mentioned in the first post t-raid was not included in FlexRAID Standards. So does this statement mean FlexRAID Standards will perform better than t-raid? (I'm not being a smart-ass, I really want to know...)
That comparison was against standard RAID implementations and not against tRAID or RAID-F.
Transparent RAID has many features you won't find in FlexRAID Standards™.
For most users, I find Transparent RAID to be the better RAID system over standard RAID implementations including FlexRAID Standards™.
As standard RAID tends to stripe data for performance, yes, the performance of standard RAID will typically be better. In exchange for that performance though, you will lose a lot of flexibility such as:
- if you lose a disk beyond the tolerance level, you will lose ALL your data on the RAID
- there is no support for disks with existing data except for RAID 1
- you cannot access the data on the individual disks as the data is striped
- you lose energy efficiency as all the disks are spun during reads and writes operations
- no easy RAID Expansion or Contraction (though FlexRAID Standards™ does provide a RAID migration feature)
RAID-F and tRAID were created because there is a lot of limitations with standard RAID with the only benefit being better performance.
For the typical home user, RAID-F and tRAID provide all the performance needed with none of the limitations.
If accessing the storage from the network, RAID-F or tRAID + Landing Disk will typically saturate a Gigabit connection. So, having a particularly fast back-end storage here makes little difference.
FlexRAID Standards™ is being released because there are still good use cases for it (all of them relate to performance needs):
- Faster video editing storage
- Faster database server storage
- Faster web server storage
- Storage for high concurrency workloads
- Etc.
Outside of RAID 1 (which will be included in tRAID btw), the other RAID implementations are mostly of interest to businesses with heavy workloads or to home users that like nice performance benchmarks.
