Can disks in RAID 5 seek independently?

Hi All

I'm trying to understand how I/O size affects throughput and bandwidth for RAID Groups on a CX4-120 SAN

I have the following queries:

a) Can the heads on disks in a RAID 5 Group seek independently or do they always have to read/write the same stripe?

Wikipedia states that for RAID 0 "For reads and writes that are smaller than the stripe size, such as database access, the drives will be able to seek independently" but is not clear for RAID 5.

http://en.wikipedia.org/wiki/Standard_RAID_levels

b) Will the I/O Size that the processors receive be the same I/O size that the processors send to the RAID Groups?

What I mean to say, is that will the processors amend the I/O requests to optimise performance on the RAID Groups? If so, does anybody know how and if there is documentation for this?

I have read about the scenario where the processor prefetches to cache for sequential writes and also will coalesce writes in cache to (I presume) be sequential and stripe sized.

One example I have in mind is read cache misses, will the processor alert the I/O request or would it always send it on, unaltered?

Thanks

John